INTERNATIONAL JOURNAL OF CLIMATOLOGY Int Climatol
53K - views

INTERNATIONAL JOURNAL OF CLIMATOLOGY Int Climatol

COMRIE BRENT FRAKES and DAVID P BROWN Department of Geography En ironment Institute Pennsyl ania State Uni ersity Uni ersity Park PA USA Department of Geography and Regional De elopment Uni ersity of Arizona Tucson AZ USA Department of Geography So

Download Pdf

INTERNATIONAL JOURNAL OF CLIMATOLOGY Int Climatol




Download Pdf - The PPT/PDF document "INTERNATIONAL JOURNAL OF CLIMATOLOGY Int..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.



Presentation on theme: "INTERNATIONAL JOURNAL OF CLIMATOLOGY Int Climatol"— Presentation transcript:


Page 1
INTERNATIONAL JOURNAL OF CLIMATOLOGY Int Climatol 21 : 1923–1950 (2001) REVIEW DEVELOPMENTS AND PROSPECTS IN SYNOPTIC CLIMATOLOGY BRENT YARNAL a, *, ANDREW C. COMRIE , BRENT FRAKES and DAVID P. BROWN Department of Geography En ironment Institute Pennsyl ania State Uni ersity Uni ersity Park PA USA Department of Geography and Regional De elopment Uni ersity of Arizona Tucson AZ USA Department of Geography Southern Illinois Uni ersity Carbondale IL USA Department of Geography and Regional De elopment Uni ersity of Arizona Tucson AZ USA Recei ed May 2000 Re ised 24 March 2001

Accepted 26 March 2001 ABSTRACT Developments in synoptic climatology in the 1990s included advances in traditional synoptic climatology, empirical downscaling, and dynamical downscaling (i.e. regional climate modelling). The research emphasis in traditional, empirical–statistical approaches to synoptic climatology shifted from methodological development to applications of widely accepted classification techniques, including manual, correlation-based, eigenvector-based, compositing and indexing schemes. In contrast, most efforts in empirical downscaling, which became a well-established field of

synoptic climatology during the 1990s, were directed to model development; applications were of secondary concern. Similarly, regional climate models (RCMs) burst onto the scene during the decade and focused on model development, although important progress was made in linking or coupling RCMs to regional or local surface climate systems. This paper discusses prospects for the future of traditional synoptic climatology, empirical downscaling and regional climate modelling. It concludes by looking at the present role of geographic information system (GIS) concepts in synoptic climatology and

the potential future role of GIS to the field. Copyright  2001 Royal Meteorological Society. KEY WORDS : climate downscaling; regional climate modelling; synoptic climatology DOI: 10.1002 joc.675 1. INTRODUCTION At the end of his review of methods and applications in synoptic climatology, Yarnal (1993) considered the future of the field. He speculated that methodological development would go in four directions. One would be the continued fine-tuning of the existing empirical–statistical approaches to synoptic climatology. Another would involve new empirical–statistical techniques, such

as neural network analysis, that provided the tightest possible mapping of the complex, non-linear relationships between the atmosphere and the surface environment. A third would include the use of simulation models to link the cascade of physical scales and processes from the large-scale atmospheric circulation to the surface environment. Finally, he speculated that geographic information systems (GIS) would be a critical link between the atmosphere and the surface. During the 1990s, three of these four areas were major methodological thrusts in synoptic climatology. Traditional synoptic

climatology maintained a steady following as methods continued to improve. Empirical downscaling, which evolved from early synoptic climatological efforts in specification (see Yarnal, 1993, chapter 5), captured the attention of a significant proportion of the field and attracted many scientists from outside synoptic climatology. Dynamical downscaling, more commonly known as * Correspondence to: Department of Geography Environment Institute, Pennsylvania State University, University Park, PA 16802, USA; e-mail: alibar@essc.psu.edu Copyright  2001 Royal Meteorological Society
Page

2
B. YARNAL ET AL. 1924 regional climate modelling, added dynamical numerical modelling to the traditional empirical statistical modelling of synoptic climatologists. Only GIS failed to develop as a significant component of synoptic climatology. This paper will review developments in these three major areas of synoptic climatology during the 1990s. Section 2 will address work in traditional synoptic climatology, while Section 3 will survey empirical downscaling. Section 4 will examine regional climate modelling. The paper will conclude by contemplating prospects for synoptic climatology

in the first decade of the new century. Because this review expands commonly held notions of synoptic climatology, it is important to ensure that the definition of synoptic climatology encompasses this expansion. We will save that task to the last section of the paper, but offer two working definitions of synoptic climatology proffered by Yarnal (1993, p. 5) that will enable us to address the problem. First, he suggested simply that synoptic climatology relates the atmospheric circulation to the surface environment . Then, he refined that idea by saying synoptic climatology integrates the

simultaneous atmospheric dynamics and coupled response of the surface environment . We will refer to these definitions as Definition I and Definition II, and will discuss them briefly at the end of the three following sections. 2. DEVELOPMENTS IN TRADITIONAL SYNOPTIC CLIMATOLOGY Until the late 1980s, it was not uncommon to find papers in the synoptic climatology literature that emphasized a classification scheme, while limiting discussion of its application. Reviewers and editors saw to it that this practice stopped, and that the classification was used as a means to an end, scientifically

speaking. The use of synoptic climatologies has since become sufficiently mainstream so that the technique sometimes does not appear in the title or keywords of a paper. This advance likely reflects widespread acceptance of the techniques, as well as a maturing of the available approaches to the point where they may be applied in a standard way. The trend is most evident in the compositing and superposed epoch analysis literature, but examples exist for all synoptic classifications. This section examines the traditional synoptic climatological literature of the 1990s using the taxonomy of

classification techniques proposed by Yarnal (1993). These techniques have become so widely accepted that it is impractical to document exhaustively in this space the published literature of the past decade; instead, the paper highlights major themes and significant recent contributions to the field. Manual, correlation-based, eigenvector-based, compositing and indexing schemes are all considered here, along with the more recently introduced techniques of spatial synoptic classification and synoptic dendroclimatology. In many instances, authors have used more than one technique in their

respective analyses, and we attempt to note this accordingly. While most of the work done in synoptic climatology in the last 10 years comprises regional case studies, some analyses of the techniques themselves have been undertaken (e.g. El-Kadi and Smithson, 1992; Yarnal, 1993). These analyses are also identified with their appropriate techniques. In addition, the section focuses on the typical themes of study, such as air quality, storm frequency etc., which are evident in the bodies of work of the various classification techniques. 2.1. Manual typing Subjective manual classification of

synoptic circulation patterns has long been a mainstay of synoptic climatological methodology. Manual typing is not easily replicated and requires a labour-intensive endeavour on the part of the researcher, but at the same time, the method allows the informed analyst far greater insight into climatic subtleties that might otherwise be missed. Two papers detailed possible improvements to the manual typing technique. Comrie (1992) identified a procedure for removing the synoptic climate signal from environmental data ( declimatizing ) by using weather-type frequencies. These frequencies can

assist in discriminating between within-type and between-type variations in climate-related environmental time series data. A second paper (Frakes and Yarnal, 1997) recognized the Copyright  2001 Royal Meteorological Society Int Climatol 21 : 1923 1950 (2001)
Page 3
SYNOPTIC CLIMATOLOGY: DEVELOPMENT AND PROSPECTS 1925 main drawbacks of the manual classification technique as being highly subjective and quite labour-intensive. They proposed a simple procedure that combined the manual classification and correlation-based classifications (discussed in the next section) to

minimize the weaknesses of both methods, discovering that the new hybrid procedure was at least comparable with standard manual classification at daily time scales. Many of the regionalized manual classification studies of the past decade were focused on Pennsylvania. Comrie (1990, 1994) first considered the synoptic climatology of surface ozone in rural areas of the state, and his further work continued manual classification of ozone pollution in the metropolitan region of Pittsburgh (Comrie, 1992; Comrie and Yarnal, 1992). Yarnal and Draves (1993) used manual analyses to create a synoptic

climatology of stream flow and acidity in central Pennsylvania. Yarnal and Frakes (1997) and Lakhtakia et al . (1998) used synoptic climatology in analyses of discharge events of the Susquehanna River Basin and its hydrographic response to a single storm event. In addition, Zelenka (1997) manually identified seven synoptic types for the summer months in a study aimed at linking meteorological patterns with acid aerosol concentrations in Uniontown. Outside of Pennsylvania, synoptic climatologies of snowfall and snowpacks were popular subjects for manual classification. Leathers and Ellis (1996)

linked synoptic mechanisms with snowfall increases around Lake Erie and Lake Ontario, and Grundstein and Leathers (1999) considered snow-surface energy exchanges in the northern Great Plains region as they relate to forcing mechanisms at the synoptic scale. In addition, Neale and Fitzharris (1997) examined the energy balance of a snowpack in the New Zealand Alps. Air quality studies were also conducted using manual synoptic classification. Davis and Gay (1993a) used this technique to examine air quality in Grand Canyon National Park as a function of synoptic patterns. Gatebe et al . (1999)

developed a synoptic classification to determine significant months for use in an air transport climatology for Kenya. Meteorological parameters were themselves analysed in several manual classification papers during the 1990s. Sweeney and O Hare (1992) linked daily precipitation distribution across Britain and Ireland with Lamb-classification circulation types. (Use of the Lamb classification was terminated upon Professor Lamb s death and was replaced by an automated method that seeks to mimic the original; the two techniques were compared by Jones et al ., 1993.) Davis et al . (1993) created

a synoptic climatology of Atlantic coast northeasters by identifying eight distinct storm types through manual classification. Marroquin et al . (1995) used a subjective method to classify days belonging to rainfall generating mechanisms (RGMs) based on 500 hPa patterns and surface topography. Weber (1998) created a synoptic climatology of wind flow patterns around Basel, Switzerland, and Prezerakos (1998) classified synoptic circulation patterns associated with temperature inversions over Athens, Greece, that persisted for more than 24 h. 2.2. Correlation based analyses The popularity of

correlation-based map-pattern classification springs from its intuitive and simple basis of automating the same task performed manually by an analyst (Yarnal, 1993). Its product is easily read and understood by the user, although it may lack some of the conciseness of the eigenvector analyses (see Subsection 2.3). The Lund and Kirchhofer classification schemes have been staples of correlation-based analyses since their inception, owing to their automated nature and consequent decreased emphasis on researcher subjectivity. The Kirchhofer technique has become especially popular in recent years,

with El-Kadi and Smithson (1992) identifying it as the most advantageous synoptic classification methodology. However, Blair (1998) noted two problems with the Kirchhofer scheme and proposed a corrected algorithm to compensate for the errors, and Kaufmann et al . (1999) used a Monte Carlo approach to examine the significance of synoptic patterns generated through use of the Kirchhofer technique. Classifications of precipitation patterns often make use of correlation-based analyses. Knappenberger and Michaels (1993) used canonical correlation analyses (CCA) to relate cyclone frequency to

precipitation and temperature at the surface for the mid-Atlantic region. Konrad (1997) used automated Copyright  2001 Royal Meteorological Society Int Climatol 21 : 1923 1950 (2001)
Page 4
B. YARNAL ET AL. 1926 typing to classify heavy rainfall events based on synoptic regime characteristics for the southeastern United States. Connor and Bonell (1998) linked trade wind precipitation on the northeast Queensland, Australia, coast with synoptic patterns, while Ruiz and Vargas (1998) used upper-level vorticity analyses to distinguish precipitation distribution over Argentina. At

a much broader spatial scale, Saunders and Byrne (1999) used a Kirchhofer methodology with 500 hPa geopotential heights to create grid-scale precipitation data for large portions of North America and the Pacific Ocean. Temperature variability was also a topic of focus in the correlation-based literature. Levey (1996) correlated atmospheric circulation with temperature variability at Cape Town, South Africa. Konrad (1998) related planetary-scale circulation anomalies to cold air outbreaks in the southeastern United States. Most recently, Brinkmann (1999a) used correlation-based classification

to remove sources of within-type variability of winter temperatures over the Lake Superior basin. Other less obvious research subjects made use of correlation-based synoptic climatologies. Brook et al (1995) undertook a dual synoptic and chemical climatology of eastern North America, using correlations to relate precipitation and meteorological category to sulphate, nitrate and acidity deposition. An additional study was that of Marshall et al . (1998) who correlated synoptic-scale climatic variations with electrical conductivity traces recorded in firn cores from the Antarctic Peninsula. 2.3.

Eigen ector based analyses The advent of mainframe and desktop computers now permits the myriad calculations necessary for eigenvector-based synoptic climatologies. The mathematical compactness and statistical robustness of these methods makes them popular in the research community. Many contributions to the literature were made during the 1990s. The eigenvector-based classifications (Figure 1) of the past decade relied mainly on two related techniques, principal components analysis (PCA) and empirical orthogonal functions (EOFs). Yarnal (1993) identified the subtle differences between PCA and

EOF approaches; thus, because of the high number of citations using these two techniques, we treat reviews of PCA- and EOF-based studies separately. Other studies have made individual use of the multivariate statistical techniques of cluster analysis, discriminant analysis and CCA; these papers are discussed as well. 2.3.1. PCA . Principal components analysis tends to convey more information than EOFs do, and more PCA software for both mainframe and desktop use exists. Consequently, PCA analysis has long been a favourite of climatologists working outside of the meteorology discipline (Yarnal,

1993). Davis and Kalkstein (1990), who created an automated spatial index of synoptic situations, proposed variations on standard PCA-based classifications. Bloomfield and Davis (1994) used the orthogonal rotation of principal components to increase interpretability of climate data, and Lana and Fernandez Mills (1994) deduced the minimum sample size needed to obtain reliable classifications using PCA. Chan and Jiu-En (1997) created projection-pursuit PCA and showed it more robust than traditional EOF classifications. In addition, Yao (1998) employed orthogonal transformations between principal

components and the original variates of the data to create a loading correlation model suitable for use as a classification technique. Synoptic climatologies of rainfall patterns frequently made use of PCA. Lyons and Bonell (1994) compared several variations of PCA in a study of daily rainfall over Townsville, Queensland. van Regenmortel (1995) used PCA to create a 9-year realization of Botswana rainfall patterns. Sumner et al (1995) extracted principal components from surface circulation patterns to establish a linkage to daily rainfall for the Mallorca region of the Mediterranean. Serra et

al . (1998) adopted the -mode PCA in a synoptic climatology of daily precipitation for Catalonia, and Romero et al . (1999) classified synoptic patterns based on daily rainfall for the entire Spanish Mediterranean area. Air quality studies often make use of synoptic climatologies, and there were several examples in the recent literature of PCA-based analyses. Davis and Gay (1993b) used an upper-air synoptic climatology, based on a combination of PCA and cluster analysis, to assess air quality variations over the southwestern United States; similarly, Comrie (1996) created an all-season

synoptic climatology of air pollution, concentrated at the United States Mexico border region. McGregor and Bamzelis (1995) used both PCA Copyright  2001 Royal Meteorological Society Int Climatol 21 : 1923 1950 (2001)
Page 5
SYNOPTIC CLIMATOLOGY: DEVELOPMENT AND PROSPECTS 1927 and cluster analysis in their air pollution study for Birmingham, UK. Shahgedanova et al . (1998) defined a synoptic climatology of Moscow air pollution, and Avila and Alarcon (1999) conducted a precipitation chemistry study for a rural location in northeastern Spain. Among studies concentrating on

severe events, Holt (1999) used a variation on PCA, principal factor analysis, to conduct a synoptic climatology of extreme surge events off the coast of western Europe. Another severe weather-based study was that of Davis and Rogers (1992), who created a synoptic climatology of severe storms in Virginia. Davis et al . (1997) followed suit with a more focused synoptic climatology of Virginia tornadoes. PCA was also used in several more meteorologically focused analyses. Davis et al . (1991) employed PCA to analyse sea-level pressure (SLP) over the eastern United States and Davis and Benkovic

(1994) examined the temporal and spatial variations of the Arctic circumpolar vortex during January. Mote (1998a,b) determined common synoptic patterns for the Greenland ice sheet, while Brinkmann (1999b) applied an -mode PCA to 700 hPa heights over the Lake Superior basin during winter. Figure 1. Eigenvector-based classification in synoptic climatology (modified from Yarnal, 1993, reproduced by permission of Belhaven Press). Copyright  2001 Royal Meteorological Society Int Climatol 21 : 1923 1950 (2001)
Page 6
B. YARNAL ET AL. 1928 2.3.2. EOFs . While EOFs present no obvious

advantage over PCA, meteorologists have long used them because of their introduction by Lorenz (1956). The variability of precipitation was the focus of several papers employing EOF analysis. Chu and He (1994) used CCA in conjunction with EOFs in long-range predictive studies of Hawaiian winter rainfall. A study by Cacciamani et al . (1994) focused on the mesoclimatology of both winter precipitation and temperature for the Po Valley using a monthly EOF analysis. North African rainfall variability on decadal and interannual time scales was the subject of a paper by Moron (1997), and

Rodriguez-Puebla et al . (1998) characterized Iberian Peninsula precipitation variability through the use of EOFs. Lenters and Cook (1999) also examined precipitation variability, this time for the summer months and encompassing the whole of South America. The Asian monsoon was a popular subject as well, treated in turn by Kripalani and Kulkarni (1998), Kang et al . (1999), Singh (1999) and Svensson (1999). A variety of research topics was present in other EOF literature of the 1990s. Kidson (1994a, 1997) applied EOFs to the study of daily and monthly weather patterns in New Zealand; following

suit, White and Cherry (1999) examined the influence of the Antarctic Circumpolar Wave on New Zealand temperature and precipitation during the fall and winter seasons. Kidson and Watterson (1995) used output from the Australian Commonwealth Scientific and Industrial Research Organisation (CSIRO) model to analyse cyclonic climatology off the New Zealand coast. Walland and Simmonds (1997) associated synoptic circulation patterns with snow cover variability for both North America and Eurasia. Roswintiarti et al . (1998) determined that local scale teleconnections to large-scale events such as El

Nin could be linked using EOFs and CCA. 2.3.3. Other multi ariate classifications . Some researchers made cluster analysis the method of choice for their synoptic climatologies. Fernau and Samson (1990) applied this methodology to transport vectors in eastern North America; Dorling et al . (1992a,b) applied it to air trajectory data in order to better understand pollution climatology in southern Scotland. Kidson (1994b) furthered his work in the New Zealand region on daily climatic data, and Sumner (1996) examined daily precipitation patterns over Wales, England using cluster analysis and the

Lamb classification. Still another paper (Dilley, 1996) examined daily precipitation patterns for the summer season in Oaxaca, Mexico. Diab et al . (1991) identified the synoptic types associated with rainfall using discriminant analysis. 2.4. Compositing Compositing as a synoptic classification method involves the averaging of a set of maps that meet some specified criterion or criteria. It is a useful technique for taking a first cut at understanding a climatic dataset (Yarnal, 1993). In disciplines outside of geography, such as in tree-ring research, this method is sometimes called

superposed epoch analysis and is often used with indexing (e.g. D Arrigo et al ., 1993). Studies based in the winter season enjoyed some popularity in the literature. Rutlant and Fuenzalida (1991) provide an example of compositing by averaging 500 hPa contour anomaly fields in a study of Chilean winter rainfall. Garfin (1998) established linkages between tree ring anomalies in the Sierra Nevada and winter circulation patterns. Hartley and Keables (1998) associated snowfall variability in New England with composited atmospheric circulation patterns. A focus on the atmosphere itself was present

in two papers using compositing. Harvey and Hitchman (1996) established an Aleutian High Composite in their investigation of the climatological properties of the persistent stratospheric high-pressure system over the Aleutian Islands during winter. Schubert et al (1998) used compositing in an analysis of warm-season moisture flows over the central and eastern United States. Other compositing studies included the extreme event analyses of Leathers (1993), who created a synoptic climatology of tornadoes in the northeast United States, and of Harnack et al . (1998), who focused their research on

extreme summer rainfall events in Utah. Mock et al . (1998) outlined the use of composite maps of circulation, temperature and precipitation to describe the variability of climatic elements at the surface. Copyright  2001 Royal Meteorological Society Int Climatol 21 : 1923 1950 (2001)
Page 7
SYNOPTIC CLIMATOLOGY: DEVELOPMENT AND PROSPECTS 1929 2.5. Indexing The use of indexing is commonly associated with synoptic climatological studies attempting to show linkage to large-scale climatic events. Often these linkages involve hemispheric- or global-scale phenomena, such as El Nin

Southern Oscillation (ENSO), North Atlantic Oscillation (NAO), or Pacific Decadal Oscillation (PDO) fluctuations. Synoptic patterns can be indexed to facilitate correlation analyses with other climatic data. Simple indices of circulation, when handled properly, can encapsulate much of the information captured by more elaborate classification techniques (Yarnal, 1993). Examples of the work done on the linkages of atmospheric circulation to ENSO and NAO can be found in the synoptic climatological literature. Speer and Leslie (1997) matched frequencies of Australian ridging events with ENSO

fluctuations, while Dilley (1997) noted how maize yields in Mexico are affected by ENSO. Stefanicki et al . (1998) related the NAO to changes in high-pressure weather-types, and Chen and Hellstrom (1999) analysed the relationship between Swedish temperature variations and the NAO. Again, rainfall variability was a major topic for studies using the indexing technique. Lyons and Bonell (1992) created indices of rainfall activity for different synoptic circulations over Queensland, Australia, in addition to producing composite correlation field maps. Inamdar and Singh (1993) were able to utilize

regression equations to demonstrate relationships between large-scale synoptic indices and 700 hPa contour heights for the Asian monsoon. Kutiel et al . (1996) created circulation indices for the Mediterranean region and Europe, and linked them to rainfall conditions throughout the Mediterranean. Spellman (1997) evaluated the effectiveness of the commonly used zonal index in accounting for precipitation variations in England and Wales, and Bonsal et al . (1999) associated atmospheric circulation with Canadian precipitation variability. Other studies found in the literature of the past decade

included Woodhouse (1997), who created synoptic indices in an examination of winter climate and circulation for the Sonoran Desert region, and Greene et al . (1999), who constructed a synoptic climatology of air pollution around the United States by using a synoptic index. 2.6. New classification techniques Three new synoptic climatological methodologies were introduced into the literature during the 1990s. One of these is spatial synoptic classification (SSC), which is optimized for continental-scale studies using air mass characteristics. Greene (1996) used the methodology in a synoptic

climatology of summer precipitation in the eastern United States. Kalkstein et al . (1996) developed air mass frequencies for the winter season east of the Rocky Mountains, and Kalkstein et al . (1998) continued the use of SSC in a study of character and frequency changes in United States air masses. Another innovative methodology recently put forth is that of synoptic dendroclimatology (Hirschboeck et al ., 1996). This new sub-discipline of both synoptic climatology and dendroclimatology links the two fields by using tree rings to reconstruct past climate from the perspective of atmospheric

circulation. There is great potential for the use of synoptic dendroclimatology in climatic, archaeological and other interdisciplinary studies. The third new methodology uses self-organizing maps (SOMs; Kohonen, 2001) for synoptic categorization. This approach developed from neural network approaches to downscaling (see Subsection 3.2.3 below) as an analytical phase that precedes the empirical modelling. Its comparative advantage lies in the non-linear aspect of its classification; however, the choice of the number of categories is arguably even more arbitrary than in other multivariate

classification methods. Cavazos (1999) has analysed extreme precipitation events in northeastern Mexico Texas, and in the Balkans using SOM techniques. 2.7. Conclusions Traditional approaches to synoptic climatology remain popular and useful. These approaches clearly support Definition I of synoptic climatology because they relate the atmospheric circulation to the surface environment. Nevertheless, because this relationship is empirical, the physical mechanisms that link the atmosphere to the environment must be inferred and cannot be certain. Thus, traditional synoptic Copyright 

2001 Royal Meteorological Society Int Climatol 21 : 1923 1950 (2001)
Page 8
B. YARNAL ET AL. 1930 climatology only implicitly integrates the simultaneous atmospheric dynamics and coupled response of the surface environment (Definition II). The benefit of these synoptic classifications is as a tool for climate diagnostics. They aid not only studies of how a particular synoptic regime works over a given region, but also in the analysis of frequency, trends, intensity and more. Traditional synoptic climatological techniques have at times been used as forecasting tools, but because they

use a small set of generalized discrete weather-types, they are generally bettered by other empirical and numerical predictive approaches. Although the holistic approach is one of the main benefits of synoptic climatology, the multivariate synoptic classification schemes have generally stayed within a limited set of variables, and have not integrated a large variety of moisture, vorticity, pressure and other data simultaneously. Part of the problem has been the lack of datasets enabling easy access to a large range of sophisticated variables, but the emergence of reanalysis data (Kalnay et al

., 1996), for example, has begun to alleviate this concern. The general ease with which these and future datasets may be accessed is likely to stimulate activity in this regard. 3. DEVELOPMENTS IN EMPIRICAL DOWNSCALING Two emphases of empirical downscaling distinguish it from conventional synoptic climatology. First, the traditional goal of synoptic climatology is to understand the relationships between the atmospheric circulation and the surface environment. Empirical downscaling, however, focuses on trying to describe the relationships precisely, devoting less attention to physical

interpretation and more attention to model parsimony and accuracy. A second, related emphasis of empirical downscaling is its use in predicting contemporary regional and local climates from large-scale circulation data and in developing regional and local scenarios for studying the impacts of climate change. In the latter case, with the development of general circulation models (GCMs) came the realization of their inability to represent sub-grid, regional- and local-scale climate information. Although the long-term goal of numerical climate modellers is to develop accurate fine-grained model

scenarios, it became obvious in the late 1980s that this goal might take decades to realize. Because climate impact assessors needed believable regional- and local-scale scenarios immediately, empirical downscaling became a workable alternative. Consequently, empirical downscaling emerged as a stopgap effort in the late 1980s, and took on a life of its own in the 1990s. (Regional climate models (RCMs), the focus of Section 4, also emerged at this time for the same reason.) Despite the assumption that empirical downscaling could provide local scenarios from large-scale data at once, and as

reported in Section 2 unlike traditional synoptic climatology in the 1990s, empirical downscaling has focused almost exclusively on methodological concerns. Similar to its predecessor, specification (see chapter 5 of Yarnal, 1993), empirical downscaling starts by developing a mathematical relationship between contemporary observed (or analysed) large-scale atmospheric variables and observed (or analysed) elements of the sub-grid local surface environment. The ability of the GCM to predict contemporary local climate is then tested by using that transfer function with a GCM simulation of the

present day. If the prediction is satisfactory, then the assumption is made that the present synoptic climatological relationship will persist into the future even with climate change and the transfer function is used to project future surface conditions from future atmospheric states simulated by the GCM. The remainder of this section will review empirical downscaling, including its major assumptions and general procedures. Because of the explosive growth of empirical downscaling during the 1990s, a comprehensive discussion of all publications is beyond the scope of this review. Consequently,

we emphasize recent work here. See Cohen (1990), Giorgi and Mearns (1991), Robock et al . (1993), Wilby (1995), Hewitson and Crane (1996), Joubert and Hewitson (1997), Wilby and Wigley (1997) and Xu (1999) for earlier empirical downscaling discussions and reviews. Copyright  2001 Royal Meteorological Society Int Climatol 21 : 1923 1950 (2001)
Page 9
SYNOPTIC CLIMATOLOGY: DEVELOPMENT AND PROSPECTS 1931 3.1. Empirical downscaling assumptions Empirical downscaling shares the assumptions common to all synoptic climatological studies (Yarnal, 1993). Two additional assumptions of

importance to empirical downscaling include stability of the atmosphere surface relationships over time and integrity of GCM output (Hewitson and Crane, 1996; Winkler et al ., 1997; Easterling, 1999). Empirical downscaling assumes that relationships between the atmosphere and environment are time-invariant. Yet, atmosphere environment relationships can be unstable for many reasons. The least probable explanation for transient relationships is a change in the underlying physics as climate changes. A more likely reason is that short-term relationships are conditional on longer-term variations in

the climate system. Unstable relationships can be enhanced further when there are multiple factors responsible for a surface condition, each of which can change in their own way. Precipitation, for instance, is controlled not only by vertical ascent, but also by available atmospheric moisture. That numerous studies have documented the instability of relationships (e.g. Wilby, 1994; Heyen et al ., 1996; Huth, 1997; Wilby, 1997) raises serious questions about the long-term reliability of the empirical downscaling models. Only a few studies (e.g. Solman and Nun ez, 1999) have demonstrated that

relatively stable models can be developed. Empirical downscaling assumes that GCMs adequately represent the large-scale features of the atmosphere. Inherent in this assumption is that general circulation modellers have an adequate understanding of atmospheric physics and that the free atmosphere is not dominated by poorly resolved boundary conditions (Winkler et al ., 1997). GCMs, however, show considerable error, even when modelling the largest-scale circulation features (e.g. Heyen et al ., 1996; Wilby et al ., 1998a; Easterling, 1999), although the upper levels are more accurately

represented than the surface (Hewitson and Crane, 1996; Sailor and Li, 1999). For instance, Martin et al . (1997), Palutikof et al . (1997) and Sailor and Li (1999), respectively, demonstrated that problems in GCM simulations of long-wave amplitudes, pressure fields and geopotential heights, and interannual variability of the free atmosphere resulted in empirical downscaling errors. Results from empirical models are neither better nor worse than those from physically based numerical models. On the one hand, empirical models suffer from a lack of clearly defined physical mechanisms. Moreover,

the empirical relationships derived by these models are inconstant over time because internal and external forcing factors vary. On the other hand, although physical relationships are clear and remain stationary in numerical models, because these models represent closed systems, they cannot account for external transient influences, which can affect their long-term accuracy. Modellers parameterize poorly-understood or difficult to model mechanisms, which weakens model output. In the end, empirical and numerical models both have strengths and weaknesses. In neither case do the weaknesses

outweigh the strengths. To conclude that one is better is erroneous each type of model provides a complementary perspective on the climate system. 3.2. Empirical downscaling procedures While numerous empirical downscaling methods exist, most studies typically share a common set of procedures (Robock et al ., 1993; von Storch et al ., 1993; Gyalistras et al ., 1994; Frey-Buness et al ., 1995; Hewitson and Crane, 1996; Easterling, 1999). This section reviews each of the main steps in Figure 2, highlighting common empirical downscaling techniques, as well as important developments. This diagram

mimics the procedure diagrams used in Yarnal (1993) and represents the inclusion of empirical downscaling as a standard synoptic climatological technique. 3.2.1. Surface ariable and local domain selection . The first step in empirical downscaling is the selection of the surface variable of interest and of the local domain in which it is located. The variable must be measurable, quantifiable and presumed to be controlled by the large-scale circulation. Important decisions must be made about the time scale of the surface data and of the analysis. Copyright  2001 Royal Meteorological

Society Int Climatol 21 : 1923 1950 (2001)
Page 10
B. YARNAL ET AL. 1932 Figure 2. Typical steps in an empirical downscaling Many empirical downscaling studies have focused on temperature (e.g. Schubert and Henderson-Sellers, 1997; Winkler et al ., 1997; Sailor and Li, 1999). Temperature is continuous, normally distributed and, therefore, requires minimal transformation, although Winkler et al . (1997) transformed temperature by the log value. Model explanation is usually good and ranges between 70 and 90% for any season, although explained variance of temperature is best for winter

months (Schubert and Henderson-Sellers, 1997; Sailor and Li, 1999; Solman and Nun ez, 1999). Precipitation is also frequently modelled in empirical downscaling studies (e.g. Corte-Real et al ., 1995; Conway and Jones, 1998; Wilby et al ., 1998b). Nonetheless, precipitation is difficult to downscale because it is discrete and often controlled by local and mesoscale processes. Precipitation also exhibits a skewed distribution and zero lower bound. Consequently, when parametric empirical downscaling models are used, precipitation must be transformed to fit a normal distribution (e.g. Huth, 1997).

While fitting a variable to a known distribution has been a frequent practice, Zorita and von Storch (1999) cautioned that distributions might change through time, thus leading to erroneous results when the means and variances of the variable are back-transformed. Consequently, depending on the spatiotemporal scales and season, model skill for downscaled precipitation is less than that of temperature and varies considerably ranging from 10 to 50%), being greatest at longer time scales and in austral winter (Huth, 1997; Busuioc et al ., 1999; Cavazos, 1999; Murphy, 1999). Copyright 

2001 Royal Meteorological Society Int Climatol 21 : 1923 1950 (2001)
Page 11
SYNOPTIC CLIMATOLOGY: DEVELOPMENT AND PROSPECTS 1933 In addition to the primary climate variables, other surface variables have been downscaled successfully. These variables include cloud cover (Bu rger, 1996; Enke and Spekat, 1997; Huth, 1997), sunshine (Bu rger, 1996; Enke and Spekat, 1997; Huth, 1997; Wilby et al ., 1998b), wind (Boga rdi and Matyasovszky, 1996; Bu rger, 1996; Huth, 1997), humidity (Bu rger, 1996; Enke and Spekat, 1997; Huth, 1997), sea level (Cui et al ., 1995; Heyen et al ., 1996),

pressure (Huth, 1997), evaporation (Wilby et al ., 1998b), and even flowering dates (Maak and von Storch, 1997). The time scale of the surface variable is an important consideration. Most often, analyses have focused on either daily (e.g. Charles et al ., 1999; Mearns et al ., 1999a; Wilks, 1999) or monthly (e.g. Easterling, 1999; Zorita and von Storch, 1999) data, with advantages and tradeoffs associated with each. Some analyses have addressed other time scales, ranging from the high-frequency synoptic signal (Hewitson and Crane, 1996; Huth, 1997) to low-frequency interannual variations (e.g.

Cui et al ., 1995). Empirical downscaling studies have differed in how to group the data temporally. Using one model for all seasons yields a more generalized model, uses all available information, and minimizes the risk of having unanticipated changes in timing of seasons (Winkler et al ., 1997). However, developing generalized models may mix different processes (e.g. mesoscale versus synoptic-scale precipitation) and result in compromised model skill. Analyses frequently have compiled the data by traditional seasons (e.g. Bu rger, 1996; Huth, 1997; Wilby et al ., 1998b), or have overlapped

the seasons (Sailor and Li, 1999). Still others have developed models for high-sun seasons (April September) and low-sun seasons October March (e.g. Matyasovszky and Boga rdi, 1996; Mearns et al ., 1999a), while Kidson and Thompson (1998) devised a model applicable to all seasons. Crane and Hewitson (1998) used the sine of the Julian day to represent intra-annual variations, thereby eliminating the need to develop separate seasonal models. 3.2.2. Large scale predictor and domain determination . Once the surface variable is chosen, the next step is the selection of a large-scale predictor along

with its spatial and temporal domain. Critical requirements are that the predictor controls the surface environment and that it be represented accurately by the GCM. Of all the large-scale predictors, perhaps no other is as frequently used as SLP (Hewitson and Crane, 1994; Wilby, 1994; Katz and Parlange, 1996; Schubert and Henderson-Sellers, 1997; Biau et al ., 1999; Busuioc et al ., 1999). SLP is useful because of its long record, which aids model development. Murphy (1999) found SLP to be preferable where regional-scale processes dominate, such as along seacoasts. Furthermore, Zorita and von

Storch (1999) indicated that SLP shows a relatively stable relationship with the surface environment. Geopotential heights are also frequently used. Most common is 500 hPa, a level representing mid-tropospheric circulation and storms tracks (e.g. Bu rger, 1996; Ba rdossy, 1997; Martin et al ., 1997; Winkler et al ., 1997; Crane and Hewitson, 1998; Kidson and Thompson, 1998; Weichert and Bu rger, 1998; Saunders and Byrne, 1999). Two other less frequently used levels are 850 hPa (Weichert and Bu rger, 1998; Sailor and Li, 1999) and 700 hPa (Ba rdossy, 1995). When empirical downscaling of

precipitation is the goal, it is useful to include a predictor of atmospheric moisture because changes in the hydrologic cycle are likely to be the underlying cause of future changes in precipitation (Murphy, 1999). It is important to note, however, that GCMs provide less accurate simulations of moisture than of surface pressure and geopotential heights. Some studies have used specific humidity as a predictor (e.g. Crane and Hewitson, 1998; Cavazos, 1999), while others have had success with relative humidity (Martin et al ., 1997; Easterling, 1999) and dewpoint temperature (e.g. Charles et al

., 1999). Presently, which representation is preferable for empirical downscaling remains undecided, although Charles et al . (1999) suggest that measures of relative humidity are more sensitive to changes in temperatures that would affect the saturation mixing ratios. A number of other predictors have also been used for empirical downscaling. They include 850-hPa temperature (Murphy, 1999), vorticity (Conway et al ., 1996; Wilby et al ., 1998b; Murphy, 1999), pressure gradients (Semenov and Barrow, 1997; Charles et al ., 1999), wind stress (Heyen et al ., 1996), large-scale precipitation

fields (Venugopal et al ., 1999), and variable combinations (Cavazos, 1999; Sailor and Li, 1999), including the -index (Murphy, 1999). Copyright  2001 Royal Meteorological Society Int Climatol 21 : 1923 1950 (2001)
Page 12
B. YARNAL ET AL. 1934 A few other considerations are also important for the large-scale predictor. One issue is the persistence of the variable. For instance, the incorporation of variables forward and backward in time (e.g. Winkler et al ., 1997; Crane and Hewitson, 1998; Sailor and Li, 1999) can account for the development and trajectory of synoptic-scale

systems. A second issue is data reduction through PCA (e.g. Corte-Real et al ., 1995; Schubert and Henderson-Sellers, 1997; Schubert, 1998) and CCA (e.g. Heyen et al ., 1996; Maak and von Storch, 1997; Busuioc et al ., 1999; Easterling, 1999). Data reduction is popular because it removes random noise that diminishes the surface climate association, it reduces the number of predictors and, specific to empirical downscaling, it may filter out irrelevant signals from the GCM (Heyen et al ., 1996). However, Huth (1997) warns that PCA removes information important for regional climate. Furthermore,

the stability of large-scale modes of variation which PCA identifies under a changed climate has not been addressed adequately. 3.2.3. Model selection and empirical relationship determination . The next step in empirical downscaling involves classifying the atmospheric data and determining the relationship between the atmosphere and the local environment. There are many ways to categorize the models used in empirical downscaling (e.g. Winkler et al ., 1997; Xu, 1999). Nevertheless, all fall along two axes: one representing a continuum that ranges from lumped to split, and another that ranges

from stochastic to deterministic (Figure 2). Lumping the circulation into discrete classes, or weather types, is the more traditional approach to synoptic climatology. Many techniques have been used to classify the weather-types in empirical downscaling studies, including the manual and automated versions of the Lamb weather types (e.g. Conway et al ., 1996; Conway and Jones, 1998) and the Kirchhofer classification (Saunders and Byrne, 1999). Others, including Hughes et al . (1993) and Enke and Spekat (1997) applied a modified minimum distance method that is similar to cluster analysis, while

Cavazos (1999) aggregated the large-scale circulation by using self-organizing maps. The relationship between the lumped patterns and the surface environment is then defined either deterministically or stochastically. Saunders and Byrne (1999) is one of the few studies to have used a deterministic model. A more common approach has been to define stochastic relationships between the circulation patterns and the surface environment (e.g. Martin et al ., 1997; Semenov and Barrow, 1997; Wilks, 1999). Charles et al . (1999), for instance, developed a first-order Markov chain model that represents

the probability of weather-state transitions that are determined by the large-scale circulation. The alternative to grouping the circulation into discrete groups is to consider the continuity of circulation where there are as many circulation classes as observations (Gyalistras et al ., 1994; Hewitson and Crane, 1996; Cavazos, 1999). The simplest is the linear regression model using one- or two-dimensional atmospheric fields or circulation indices. More advanced regression models have included multivariate adaptive regression splines (Corte-Real et al ., 1995) and non-linear regression, which

is a subset of artificial neural networks (Hewitson and Crane, 1996; Cavazos, 1999; Zorita and von Storch, 1999). Continuous-stochastic models have not received much attention. Semenov and Barrow (1997) considered how changes in temperature and precipitation of the GCM gridpoint changed the parameters of a weather generator. More often, studies have developed continuous-deterministic models (e.g. Easterling, 1999; Solman and Nun ez, 1999). For example, Busuioc et al . (1999) developed a linear regression equation to downscale monthly precipitation from monthly SLP. Preference for a particular

model in empirical downscaling is case specific, although discussion continues about which model is preferable in various circumstances (e.g. Schnur and Lettenmaier, 1998). Lumping circulation into discrete categories is advantageous because the patterns are assumed to remain time invariant, although their internal characteristics or frequencies may be altered under a changed climate. Continuous models have the advantage of representing the entire continuity of circulation and, because they are not generalizable, of differentiating among subtle differences in the atmospheric state that can

result in very different outcomes at the surface. It is presently unclear, however, how GCM errors propagate through such models. There is also limited agreement that continuous models are more effective than discrete models when handling continuous variables, including temperature. Furthermore, Copyright  2001 Royal Meteorological Society Int Climatol 21 : 1923 1950 (2001)
Page 13
SYNOPTIC CLIMATOLOGY: DEVELOPMENT AND PROSPECTS 1935 continuous models are thought to be more reliable for representing extreme events (e.g. Sailor and Li, 1999), which have been shown to change

through time (Karl et al ., 1996). Stochastic and deterministic models also bring their own advantages and disadvantages to empirical downscaling (Winkler et al ., 1997). Even with the relatively short integration period of most GCMs, stochastic models can generate long time series from which to develop a climatology. However, stochastic models share a number of problems, including the inability to compare predictions directly with observations and, in some cases, the inability to simulate the expected variance and autocorrelation of surface variables at the daily time step (Richardson, 1981).

Moreover, in the case of climate change, it is unclear how to adjust the stochastic parameters in a physically meaningful way (Wilby et al ., 1998a). Of the continuous-deterministic models, an important question concerns whether non-linear models are preferable to their linear counterparts. While non-linear models should be more adept in representing the non-linear relationships found in the climate system, their increased complexity and sensitivity raise doubts about their application in empirical downscaling. For instance, Schubert and Henderson-Sellers (1997), Winkler et al . (1997) and

Zorita and von Storch (1999) demonstrated that non-linear regression models, including artificial neural networks, offered little improvement over the basic linear regression models. Alternatively, others, including McGinnis (1994) and Weichert and Bu rger (1998), found non-linear models to be an improvement over the linear models in certain instances. Finally, confirmation that the empirical model is working properly is an essential step for any empirical downscaling analysis. Because relationships might not remain stable through time, it is imperative that split-sampling tests use different

periods or extreme years (e.g. Wilks, 1999). 3.2.4. GCM scenario de elopment . An essential step prior to the application of the empirical model is the evaluation of the GCM s ability to represent the large-scale predictors. When weather-types are used, studies have evaluated the similarity of type frequencies at annual and seasonal time scales, of pattern gradients, and of day-to-day persistence (e.g. Goodess and Palutikof, 1998). When continuous models are used, quantile quantile comparisons of predictors have proved insightful for detecting differences between observations and GCM results

(Karl et al ., 1990). Once the GCM integrity is confirmed, control and 2 CO integrations are run. In general, because GCMs have biases in their simulations of current climate, comparisons between the 1 CO and 2 CO models are used instead of comparisons between observations and doubled-CO simulations. 3.3. Conclusions Empirical downscaling has become an influential part of synoptic climatology in the past decade. As suggested by its prominent place in the Intergovernmental Panel on Climate Change (IPCC) Working Group I Third Assessment Report, its emphasis on prediction and projection makes it

attractive to many scientists. Yet, the inability of empirical downscaling to explain the linkages that it makes between the atmosphere and the surface environment is the approach s single biggest weakness. Empirical downscaling techniques make it difficult to infer the relationships they uncover. Thus, they tacitly, but weakly support Definitions I and II of synoptic climatology. 4. DEVELOPMENTS IN REGIONAL CLIMATE MODELLING Until the 1990s, synoptic climatology focused on developing empirical statistical associations to determine relationships between atmospheric circulations and the

environment. The field is now turning to dynamical downscaling usually called regional climate modelling as another way to link the cascade of processes from global to local. Regional climate modelling combines the traditional goal of synoptic climatology, that is, to understand the relationships between the atmospheric circulation and the surface environment, with the primary emphasis of empirical downscaling, that is, to describe the relationships precisely. Unlike empirical synoptic climatological approaches, regional climate modelling presents the scientist with the opportunity to grasp

the dynamics of individual synoptic systems and their Copyright  2001 Royal Meteorological Society Int Climatol 21 : 1923 1950 (2001)
Page 14
B. YARNAL ET AL. 1936 interaction with surface environmental systems, while characterizing the climatology of a suite of those systems and interactions. Regional climate modelling also adopts the second emphasis of empirical downscaling predicting contemporary regional climates from large-scale circulation data, and developing regional scenarios for studying the impacts of climate change. Similar to empirical downscalers, regional

climate modellers have spent more time developing and refining the models than addressing synoptic climatological relationships. Refreshingly, synoptic climatological applications are now becoming a more important part of the regional climate modelling literature. For an excellent discussion of continuing issues in RCM development, see Giorgi and Mearns (1999). This section will review the literature on a model-by-model basis, but will lump these models into a few broad categories. The section will conclude with a look at RCM intercomparisons, including studies that compare RCMs and empirical

downscaling models. See Giorgi and Mearns (1991), Giorgi (1995), Kattenberg et al . (1996), McGregor (1997) and Giorgi and Mearns (1999) for overviews of regional climate modelling from other perspectives. We do not present a procedure diagram similar to Figure 1 because the set of operator decisions are large and very different to those of empirical statistical analyses. 4.1. er iew of approaches to regional climate modelling RCMs can be viewed from top-down and bottom-up perspectives. Similar to empirical downscaling, the top-down viewpoint sees regional climate modelling as a way to improve

upon the coarse resolution of global-scale analyses of observations and of GCMs. There are two approaches to top-down regional climate modelling. In one, a modified, limited-area mesoscale meteorological model nests within a global-scale GCM. The global domain provides the lateral boundary forcing and determines the broad behaviour of the RCM. The second approach also employs nesting but, in this case, the RCM uses physical parameterizations that are consistent with those of the GCM, but tuned to account for the finer grid increments of the RCM. In either approach, because top-down RCMs

simulate regional climate on a storm-by-storm basis for seasons to decades, a certain amount of model error is inevitable. The much less common bottom-up perspective also looks at RCMs on a storm-by-storm basis, but tries to simulate each weather system with temporal and spatial precision. Initial conditions are more important than lateral boundary forcing. Spatial resolution tends to be more critical, with some experiments employing grid increments as fine as a few kilometres and consequently requiring computer-intensive non-hydrostatic model formulations. These more stringent requirements

currently limit the temporal domain of bottom-up applications to time-scales of weeks to months. It is important to note that top-down and bottom-up perspectives are not mutually exclusive, and that the two perspectives eventually will merge with improvements in model physics and computer power. 4.2. Top down RCMs One top-down approach to regional climate modelling modifies a well-tested mesoscale meteorological model originally designed to simulate individual weather systems so that the model can now accommodate simulations of these systems over weeks and longer periods. The RCM nests in a

global domain provided either by analyses of observations or by a GCM. In those cases in which the RCM nests in a GCM, there are advantages and disadvantages to this approach (Giorgi and Mearns, 1999). The principal advantage is that the parameterizations of the nested RCM and the driving GCM are optimized to their respective model resolutions. The primary disadvantage is that it is difficult to determine whether model discrepancies are scale-dependent or physics-dependent. The second approach also nests a relatively high-resolution RCM in a lower-resolution GCM. In this approach, however,

instead of the RCM being a climate formulation of a mesoscale meteorological model, the RCM is essentially the same model as the GCM. In principle, the physics and dynamics in the RCM and GCM are consistent, but the grid increments and time steps of the RCM are finer. The assumption is that GCM resolutions are coarse because of limitations in computer power, not limitations in model physics and dynamics. It is important to note that the consistent-physics approach does not Copyright  2001 Royal Meteorological Society Int Climatol 21 : 1923 1950 (2001)
Page 15
SYNOPTIC

CLIMATOLOGY: DEVELOPMENT AND PROSPECTS 1937 intend for parameterizations to be identical at all scales, but that they should be tuned to facilitate model resolution at smaller grid increments (Laprise et al ., 1998). Although some scientists dispute the idea that physics remain the same but tunable across scales, the results of simulations using this approach are comparable to those produced by the nested mesoscale model approach (Kattenberg et al ., 1996; Giorgi and Mearns, 1999). 4.2.1. RCMs deri ed from existing mesoscale models . The National Center for Atmospheric Research (NCAR) RegCM is

a top-down RCM developed by Filipo Giorgi and collaborators. It is the most-used RCM, generating more than 50 refereed publications about 40% of all papers on regional climate modelling at the time of this writing. Furthermore, approximately 60% of all refereed RCM publications involve top-down RegCMs derived from existing mesoscale models. For brevity, only a selection of these papers will be reviewed here. RegCM originated in the late 1980s (Dickinson et al ., 1989; Giorgi and Bates, 1989) with the Penn State NCAR Mesoscale Meteorological Model. Additions or modifications to the radiative

transfer, biosphere atmosphere interaction, planetary boundary layer, cumulus convection and atmospheric moisture schemes changed it from a meteorological model to one suitable for climate studies. A major upgrade in model physics and numerical schemes resulted in RegCM2 (Giorgi et al ., 1993a,b). Upgrades and modifications to RegCM2 continued throughout the 1990s (Giorgi and Mearns, 1999). Many performance and sensitivity studies, including present-day simulations (e.g. Giorgi et al ., 1993c; Liu et al ., 1994; Marinucci et al ., 1995) and climate change simulations (e.g. Giorgi et al .,

1992, 1994a; Hirakuchi and Giorgi, 1995), tested the effectiveness of these modifications to RegCM. In most of these papers, an important indicator of model performance was the ability of RegCM essentially a model of the regional atmospheric circulation to reproduce surface temperature and precipitation means, variances and spatial distributions. Thus, this research falls under the purview of synoptic climatology. Synoptic climatological applications using RegCM and going beyond temperature and precipitation have been less common. Several of these applications coupled RegCM to lake models and

showed that the model systems could realistically reproduce lake surface temperatures, evaporation rates and ice fractions and thickness (Bates et al ., 1993; Hostetler et al ., 1993; Bates et al ., 1995; Hostetler and Giorgi, 1995; Small et al ., 1999). RegCM also has been applied to basin hydrology (Giorgi et al ., 1994b) and basin water yield in a changed climate (Stonefelt et al ., 2000). Liu et al . (1996) used RegCM to simulate prolonged precipitation that resulted in region-wide floods, while Giorgi et al . (1996) utilized the model to study the contribution of local versus non-local

processes in the maintenance of droughts and floods. Seth and Giorgi (1996) induced organized mesoscale circulations akin to the sea breeze by changing soil moisture and vegetation surface cover in RegCM. In a study involving wheat yields, Mearns et al . (1997) used RegCM to help determine the importance of including changes in variance, as well as changes in mean in climate change scenarios. Finally, Mearns et al . (1999b) linked a GCM directly to two crop models and RegCM to the same models and found that model choice introduces considerable uncertainty in climate change experiments. As

implied in the preceding paragraph, an important synoptic climatological dimension of RegCM is its potential for coupling with other components of the climate system. This is an exciting development because it enables the synoptic climatologist to step beyond the black-box approach of empirical statistical studies and to address directly the physical processes that link the atmosphere to the surface environment. Giorgi (1995) built upon the regional climate modelling framework to sketch a possible coupled regional climate-system model. Other researchers accepted this challenge. One important

example based on RegCM2 is the Arctic regional climate system model (ARCSYM; Lynch et al ., 1995). To the suite of models normally coupled to RegCM2, ARCSYM adds a dynamic sea-ice model, an ocean-circulation model, and a mixed-layer ocean formulation. ARCSYM is a significant step forward in the simulation of polar climate because GCMs do a notoriously poor job in these regions. From a synoptic climatological point of view, ARCSYM is important because its climate-system approach shows inherent concern for atmosphere surface relationships. ARCSYM has been used to investigate polynya formation in

the Bering Sea (Lynch et al ., 1997), interactions between ocean circulation and polar climate Copyright  2001 Royal Meteorological Society Int Climatol 21 : 1923 1950 (2001)
Page 16
B. YARNAL ET AL. 1938 (Bailey et al ., 1997), the role of snow-albedo feedback in the onset of polar spring (Lynch et al ., 1998), and the impacts that various tundra ecosystems have on surface energy budgets (Lynch et al ., 1999a,b). Other RCMs have explored coupled or linked biogeophysical systems in the United States using variations of the Penn State NCAR Mesoscale Meteorological Model. Leung

and Ghan (1995, 1998, 1999a) designed the top-down Pacific Northwest National Laboratory (PNNL) RCM to provide more accurate simulation of precipitation in mountainous terrain than provided by GCMs and other RCMs. PNNL-RCM has been coupled to hydrologic models to study the response of mountain watersheds to contemporary climate variation (Leung et al ., 1996, 1999a) and future climate change (Leung and Ghan, 1999b; Leung and Wigmosta, 1999). Pan et al . (1999a,b) applied versions of the Penn State NCAR Mesoscale Meteorological Model to simulate flood, drought, and normal climate regimes in the

Midwest and Great Plains, especially as they relate to land-use change. Outside the United States, Lau et al . (1998) successfully simulated the evolution of the June 1994 East Asia summer monsoon using an RCM version of this meteorological model. Not all RCMs based on mesoscale meteorological formulations evolved from the Penn State NCAR model. One example is the Colorado State University Regional Atmospheric Modeling System (RAMS; Pielke et al ., 1992; Copeland et al ., 1996). RAMS is a highly flexible mesoscale model with resolutions ranging from meters to hundreds of kilometres. Despite

its popularity among meteorological applications specialists, few climatological studies have used RAMS. In contrast, the relatively little known regional climate system model (RCSM) developed at Lawrence Berkeley National Laboratory (Miller and Kim, 1996; Miller et al ., 1999) has been used in several synoptic climatological studies. The principle aim of RCSM is to link the atmosphere, surface and hydrologic system for applications such as water resource management and environmental impact assessment (e.g. Kim, 1997; Kim et al ., 1998). 4.2.2. Consistent physics RCMs . Three important

European RCMs employ the consistent-physics strategy. One such model system is part of the British Meteorological Office Unified Forecast Climate Model. In the paper that introduced the British Unified Model RCM, Jones et al . (1995) assessed control climate runs and focused on the sensitivity of the nested-model results to lateral boundary location. Following that, Jones et al . (1997) compared the responses of the GCM and the nested RCM to a doubling of carbon dioxide. The former study found that smaller domains far from the lateral boundaries provide good high-resolution simulations with a

stronger hydrologic cycle than the GCM. The latter research demonstrated strong similarities between the RCM and GCM in autumn, winter, and spring, when large-scale dynamical forcing dominates extratropical climate. In summer, when local forcing is more important, the two model responses differed substantially. In subsequent model testing, Noguer et al . (1998) compared RCM output forced by the GCM with RCM output forced by operational analyses, which allowed the authors to separate errors arising from the driving circulation and from the RCM physics. They concluded that it is necessary to

tune model physics to the scale of the model. Bhaskaran et al . (1996, 1998) also compared the British RCM and GCM in simulations of the Indian summer monsoon, with emphasis on northward-propagating intraseasonal oscillations. The RCM produced results superior to the GCM, generating spatially and temporally realistic intraseasonal oscillations that were absent in the GCM. The authors concluded that the Unified Model produced better results in the tropical experiments because of differences in regional dynamics. A consortium of European research groups developed the HIRHAM consistent-physics

nested RCM by adding the physics of the Max Planck Institute GCM (ECHAM) to the European High-Resolution Limited Area Model (HIRLAM; Dethloff et al ., 1996). The focus of HIRHAM research has been to simulate high-latitude climates, including the whole Arctic (Dethloff et al ., 1996; Rinke et al ., 1997; Rinke and Dethloff, 2000), Scandinavia (Christensen et al ., 1998) and Antarctica (van Lipzig et al ., 1998, 1999). The Scandinavian simulations included nested 19-km and 57-km RCM grids in a coupled ocean atmosphere GCM. Investigators found that the simulation of hydrologically relevant

variables, such as runoff and snow cover, improved in mountainous regions as resolution increased. The Antarctic experiments showed that the RCM overestimated some variables (e.g. vertical temperature gradients and cloud cover) and underestimated others (e.g. vertical transfer of heat and humidity), but that it was a Copyright  2001 Royal Meteorological Society Int Climatol 21 : 1923 1950 (2001)
Page 17
SYNOPTIC CLIMATOLOGY: DEVELOPMENT AND PROSPECTS 1939 substantial improvement over the GCM. An interesting application of this model is its linkage to a regional

chemistry-transport model to estimate the influence of climate change on atmospheric pollution over Europe (Langmann and Graf, 1997) and to determine the regional effects of sulphate aerosols on European radiative forcing (Langmann et al ., 1998). The German Weather Service has pursued two paths to regional climate modelling. One is more traditional and nests a modification of the EUROPA RCM in the ECHAM GCM to form the regional model (REMO; Jacob and Podzun, 1997). More effort, however, has gone into developing a nested consistent-physics RCM, with a larger-scale regional climate version of

EUROPA (Cress et al ., 1995; Podzun et al ., 1995) driving smaller, nested grids for various areas of Europe (i.e. Scandinavia, central Europe, the British Isles, and the Mediterranean). In early work, Lu thi et al . (1996) focused on determining the ability of the large-domain RCM to reproduce European climate realistically, while Scha et al . (1996) proposed a methodology for generating small-region climate change scenarios with the nested RCM. Scha et al . (1999) conducted an application of the RCM in a study of the sensitivity of precipitation to soil moisture. They demonstrated that in

any given year the soil-moisture regime could shift the border between the wet Atlantic and dry Mediterranean climates by hundreds of kilometres. Elsewhere in the world, the climate version of the Australian Division of Atmospheric Research Limited Area Model (DARLAM) also uses the consistent-physics approach of the European RCMs described above, but for Southern Hemisphere climate. Early research focused on Antarctica. Walsh and McGregor (1996) used ECMWF analyses of observations to drive the RCM over Antarctica and found the results similar to the nested GCM simulation of McGregor and Walsh

(1993). McGregor and Walsh (1994) nested a 60-km grid within the DARLAM 125-km grid to improve the delineation of contemporary and future rainfall in Tasmania. Walsh and McGregor (1995) developed a model January and July climatology for Australia and vicinity using the 125-km grid of DARLAM. Walsh and McGregor (1997) went on to simulate the observed rainfall over Australia of two contrasting years. In this case, the investigators found that although the RCM simulation was better than the GCM alone, both inadequately captured the interannual variability. Charles et al . (1999) took the unusual

step of using the output from DARLAM to drive an empirical downscaling experiment in southwestern Australia. Finally, Katzfey (1995) employed the RCM to simulate extreme New Zealand precipitation, while Renwick et al . (1998) used the model to compare contemporary and future climates in New Zealand. To improve the simulation in that mountainous country, the latter study nested a 50-km grid in the standard 125-km grid of DARLAM. The authors noted special sensitivity of the model to the land-surface specification. The Canadian RCM also employs the consistent-physics approach (Caya et al ., 1995;

Caya and Laprise, 1999). Laprise et al . (1998) made two 5-year integrations at 45-km grid increments over western Canada for 1 CO and 2 CO conditions and found the output similar to that of the driving GCM. MacKay et al . (1998) applied the RCM to the hydrological cycle of the Mackenzie Basin. They found that the RCM performed well when driven by analyses of observations. In contrast, the RCM did a poor job simulating moisture and precipitation when driven by the GCM because of the biases of the global-scale model. Thus, the RCM can only be as good as the inputs it receives. 4.2.3. Other top

down RCMs . Scientists have developed surprisingly few top-down alternatives to the RCMs based on mesoscale models and consistent physics. One promising alternative to grid-point RCMs is the spectral RCM, with which Kida et al . (1991) experimented nearly a decade ago. Following that, the National Center for Environmental Prediction (NCEP) released the regional spectral model (RSM) that nests in the NCEP Global Spectral Model (Juang and Kanamitsu, 1994). RSM has been tested for many applications (Juang, et al ., 1997), including summertime droughts and floods and wintertime ENSO impacts (Hong

and Leetmaa, 1999). Another interesting top-down option is an off-line model developed by Goyette and Laprise (1996). We would be remiss not to mention advances in variable-resolution, high-resolution GCMs (e.g. Deque et al ., 1998) that some day may make general RCMs obsolete. Copyright  2001 Royal Meteorological Society Int Climatol 21 : 1923 1950 (2001)
Page 18
B. YARNAL ET AL. 1940 4.3. Bottom up RCMs Bottom-up RCM experiments are rare because of their computational intensity. One bottom-up RCM was designed for the Susquehanna River Basin Experiment (SRBEX; Yarnal et al

., 2000), which took advantage of the variable-nesting capability of MM5 to simulate storms systems with grid increments as fine as 4 km (Lakhtakia et al ., 1998, 1999). The focus of this research was to determine the basin-scale hydrologic response to atmospheric forcing in the mid-Atlantic region (Lakhtakia et al ., 1999; Yu et al ., 1999). In a comparable bottom-up nested RCM experiment, Dudek et al . (1996) sought to determine the scale dependence of cloud-radiation interactions in the southern Great Plains. Bottom-up RCMs may never become important until teraflop computers become widely

available. At that time, top-down and bottom-up approaches will probably merge, thus making the distinction irrelevant. 4.4. Comparison studies Regional climate modelling is becoming sufficiently common that model intercomparison projects are taking place. Christensen et al . (1997) compared how well seven RCMs simulated winter and summer climate over Europe and reported relatively large, consistent biases that varied with variable and season. For instance, wintertime surface air temperature biases were negative and attributed to underestimations of incoming longwave radiation. Leung et al .

(1999b) studied the performance of three RCMs in simulating floods associated with the 1991 monsoon floods in the Yangtze River. They found that all models simulated gross flood conditions well, but possessed significant differences in their energy and hydrological cycles. The Project to Intercompare Regional Climate Simulations (PIRCS) is the largest to date, with 12 international RCM teams simulating the 1988 Midwest drought and 1993 Midwest floods. In the drought study, the RCMs simulated synoptic-scale precipitation systems well, but did a poor job reproducing mesoscale and convective

precipitation events (Takle et al ., 1999). Energy partitioning and daily temperature fields were problematic. In the preliminary flood study, most PIRCS models produced accumulated precipitation within 10% of observed totals (Arritt et al ., 2000). In a related project, Pan et al . (2000) used RegCM2 and HIRHAM to generate 50-km, 10-year climate change simulations for climate impact assessment in the United States. Although substantial differences existed between the models, the increase in spatial variability compared to the driving GCM make the simulations useful for climate impacts

scenarios. There also have been several comparisons of statistical and dynamical downscaling techniques. Kidson and Thompson (1998) concluded that linear regression was comparable with the RAMS RCM in downscaling daily temperature and precipitation in New Zealand, although RAMS was superior in regions where convective precipitation dominates. At the monthly time scale, however, the regression model downscaled precipitation more effectively. Murphy (1999) found the British Unified Model RCM superior to a regression model when downscaling monthly wintertime precipitation over Europe, although

both models produced comparable results for temperature. On the other hand, the regression model was more effective in modelling monthly summertime temperature and precipitation than the RCM simulation. In a comparative study over the United States Central Plains, Mearns et al . (1999a) demonstrated considerable differences between output from an empirical downscaling technique and from RegCM2. The RCM showed more spatial variability; generalizations on temporal variability were more difficult to make. For a selection of river basins around the United States, Gutowski et al . (2000) revealed

that an empirical method based on step-wise multiple linear regression and RegCM2 both reproduced general features of the basin climatology, but also displayed systematic biases. Their most important finding was that results are dependent on the climatology of the basin under consideration. In short, neither empirical downscaling models nor RCMs consistently outperform the other. 4.5. Conclusions Regional climate modelling offers synoptic climatologists the opportunity to go beyond empirical, black-box approaches to the field and to determine explicitly the physical mechanisms that couple the

Copyright  2001 Royal Meteorological Society Int Climatol 21 : 1923 1950 (2001)
Page 19
SYNOPTIC CLIMATOLOGY: DEVELOPMENT AND PROSPECTS 1941 dynamics of the atmospheric circulation to surface environmental processes. Thus, RCMs satisfy Definitions I and II. 5. PROSPECTS The established methods of synoptic climatological classification are now common and, consequently, are often omitted from titles and abstracts as major foci of papers. Instead, these methods provide a means to an end in a great variety of environmental studies. Many new examples and applications of the

traditional methods are evident, particularly at the regional level and in areas that have not previously utilized synoptic climatological analyses. Refinements to the existing techniques appeared in the 1990s, such as the use of multiple -means analysis in principal component-clustering studies and of Monte Carlo significance testing in correlation-based classifications. New synoptic climatological classification methodologies introduced during the last decade, such as spatial synoptic classification and synoptic dendroclimatology, have the potential to become primary techniques in future

synoptic climatologies. As computing power and availability increase, the possibilities for greater use and refinement of synoptic climatological classification methods and for the introduction of new techniques become likely. Empirical downscaling needed only one decade to become a well-established field of synoptic climatology, but there are many areas requiring further research. One is to address more directly the assumptions of empirical downscaling, i.e. stability of the atmosphere surface relationships over time and the integrity of GCM output. A second is to develop a better

understanding of the physical processes that form the bases of the empirical relationships. Another critical need is to understand the temporal and spatial scales at which atmosphere surface relationships remain stable, the points at which these relationships break down, and the ways that these relationships change over time and space. Researchers should continue to investigate differences in model performance among various empirical downscaling techniques, including why, where and when these models excel or fail. Similarly, investigators should continue to compare empirical and dynamic

downscaling, and to determine why, where and when one outperforms the other. Most importantly, while they improve methodologies, empirical downscalers must spend more time applying their transfer functions to more surface environmental contexts. Although empirical downscaling has brought high visibility to synoptic climatology, empirical downscaling has suggested to non-synoptic climatologists that synoptic climatology is a method and not a serious field of scientific inquiry. The tremendous potential of empirical downscaling, both inside and outside the field, cannot be realized until it

emphasizes application over technical development. Regional climate modelling also holds great promise, but requires continued model development and increases in computer power to reach that promise. Modellers need to develop parameterizations that are adjustable to the time and space scales required by the investigation. As in the case of empirical downscaling, this means that scientists must know at which scales relevant processes operate and at which scales cascading processes overlap. Parameterizations need to be streamlined to calculate faster and more accurately than today. RCMs must

become more flexible so that they can adjust time and space scales quickly; such adjustments presently take modelling teams weeks to years to implement. Finally, like empirical downscaling, not only must RCM development continue, but also applications of RCMs to environmental problems must increase in number and scope. Unlike empirical downscaling, however, regional climate modelling increasingly promotes the understanding of the relationships between atmospheric processes and surface processes through linked or coupled atmosphere surface models. Thus, we conclude that the three arms of

synoptic climatology reviewed in this paper traditional synoptic climatology, empirical downscaling and regional climate modelling will continue to develop, but emphasis will shift increasingly away from atmospheric classification, modelling, and simulation and towards applications to contemporary and future surface environmental problems. Further in the future, we suspect that RCM development and increased computer power will surpass, but not entirely replace, empirical downscaling as a tool for projecting future climates. We think that empirical downscaling will merge with traditional

methods to provide both better understanding and better statistical explanation of Copyright  2001 Royal Meteorological Society Int Climatol 21 : 1923 1950 (2001)
Page 20
B. YARNAL ET AL. 1942 physical processes linking the atmosphere and surface. It is also likely that as GCMs develop finer grid increments, the need for nested RCMs will diminish and eventually disappear. In sum, we do not think that traditional synoptic climatology will die because there will always be a need for first-order understanding of empirical relationships, which in turn guide simulation model

development and the comprehension of the physics linking the atmosphere to the surface. We do think that empirical downscaling models and regional climate models will evolve into flexible, high-resolution GCMs. In the introduction to this paper, we broached the possibility of expanding the definition of synoptic climatology. To this end, in closing the sections on traditional synoptic climatology, empirical downscaling and regional climate modelling, we summarized how well that area of synoptic climatology supports either Definition I or Definition II. We conclude that it would be ideal if

synoptic climatology could be redefined so that it integrates the simultaneous atmospheric dynamics and coupled response of the surface environment (Definition II). Indeed, the focus on applications, rather than methodology testifies to this aim, as does the growth of regional climate modelling. Nevertheless, because there will always be a need for synoptic climatologists to know first-order empirical relationships in the absence of complete understanding of the coupled system, then the breadth of Definition I synoptic climatology relates the atmospheric circulation to the surface environment

will continue to be useful into the foreseeable future. And what of Yarnal s (1993) speculation that GIS would become important to synoptic climatology? Although only a few studies explicitly used GIS in the 1990s, synoptic climatology benefited considerably from GIS concepts. At the broadest level, a GIS is the software and hardware used for the storage, visualization and analysis of geographic information all of which are integral components of most synoptic climatologies. The storage and distribution of atmospheric data has changed dramatically over the last decade. New data formats (such

as netCDF) have simplified data extraction, while the Internet (where transaction and storage costs are negligible) and continued governmental support (particularly from the United States) have increased access to these data. Giant leaps in the development and availability of visualization software and in its compatibility with data, such as the marrying of the grid analysis and display system (GrADS) and the netCDF format, have enabled synoptic climatologists to acquire and visualize climatic data within minutes compared with the weeks and sometimes months invested 10 years ago!

Unfortunately, the third component of a GIS analysis is still immature in some areas of synoptic climatology. In the future, however, we anticipate that advanced analytical routines will be integrated into shareware atmospheric GIS packages, including GrADS, as well as into commercial GIS software. Thus, although GIS has made few explicit inroads into synoptic climatology, its conceptual components have played an important role in the evolution of the field. Coupled with improvements in computer technology, these advances have effectively democratized empirical synoptic climatology. Not long

ago, synoptic climatology required significant human and financial resources, which often restricted research to elite institutions. Today, powerful and relatively inexpensive desktop computers, free or cheap data, and free and easy-to-use software, have levelled the playing field, at least for empirical modelling. The ease of data access, storage, and visualization are permitting most practitioners to devote their attention to the most important elements of any synoptic climatology analysis and interpretation. NOTES 1. Giorgi recently moved from NCAR to the Abdus Salam International Centre

for Theoretical Physics, Trieste, Italy, where he continues to use RegCM. 2. HIRHAM is a third-order acronym. It is a combination of the two acronyms HIR LAM (HIgh Resolution Limited Area Model) and EC HAM (ECMWF HAMburg) model. ECMWF stands for European Centre for Medium-range Weather Forecasts. 3. HIRHAM is called RACMO (Regional Atmospheric Climate Model) in some publications (e.g. van Lipzig et al ., 1998). This switching of names is a source of confusion. Copyright  2001 Royal Meteorological Society Int Climatol 21 : 1923 1950 (2001)
Page 21
SYNOPTIC CLIMATOLOGY:

DEVELOPMENT AND PROSPECTS 1943 REFERENCES Arritt RW Jr, Gutowski WJ, Takle ES, Pan Z, Anderson CJ, Otieno F, da Silva R, Caya D, Chen S-C, Christensen JH, Fox-Rabinovitz M, Giorgi F, Hong S-Y, Juang H-M, Katzfey JJ, Lapenta WM, Laprise R, Liston G, McGregor JL, Pielke RA Sr, Roads JO. 2000. Project to Intercompare Regional Climate Simulations (PIRCS): simulation of the 1993 flood over the central US. In 15 th Conference on Hydrology Long Beach CA 14 January 2000 . American Meteorological Society: Boston, MA; 25 28. Avila A, Alarcon M. 1999. Relationship between precipitation chemistry and

meteorological situations at a rural site in NE Spain. Atmospheric En ironment 33 : 1663 1677. Ba rdossy A. 1995. Stochastic downscaling of GCM-output results using atmospheric circulation patterns. In Space and Time Scale Variability and Interdependencies in Hydrological Processes , Feddes RA (ed.). Cambridge University Press: Cambridge; 119 123. Ba rdossy A. 1997. Downscaling from GCMs to local climate through stochastic linkages. Journal of En ironmental Management 49 17. Bailey DA, Lynch AH, Hedstrom KS. 1997. Impact of ocean circulation on regional polar climate simulations using the

Arctic Region Climate System Model. Annals of Glaciology 25 : 203 207. Bates GT, Giorgi F, Hostetler SW. 1993. Toward the simulation of the effects of the Great Lakes on regional climate. Monthly Weather Re iew 121 : 1373 1387. Bates GT, Hostetler SW, Giorgi F. 1995. Two-year simulation of the Great Lakes region with a coupled modeling system. Monthly Weather Re iew 123 : 1505 1522. Bhaskaran B, Jones RG, Murphy JM, Noguer M. 1996. Simulations of the Indian summer monsoon using a nested regional climate model: domain size experiments. Climate Dynamics 12 : 573 587. Bhaskaran B, Murphy JM,

Jones RG. 1998. Intraseasonal oscillation in the Indian summer monsoon simulated by global and nested regional climate models. Monthly Weather Re iew 126 : 3124 3134. Biau G, Zorita E, von Storch H, Wackernagel H. 1999. Estimation of precipitation by kriging in the EOF space of the sea level pressure. Journal of Climate 12 : 1070 1085. Blair D. 1998. The Kirchhofer technique of synoptic typing revisited. International Journal of Climatology 18 : 1625 1635. Bloomfield P, Davis JM. 1994. Orthogonal rotation of complex principal components. International Journal of Climatology 14 759 775. Boga

rdi I, Matyasovszky I. 1996. Estimating daily wind speed under climate change. Solar Energy 57 : 239 248. Bonsal BR, Zhang X, Hogg WD. 1999. Canadian prairie growing season precipitation variability and associated atmospheric circulation. Climate Research 11 : 191 208. Brinkmann WAR. 1999a. Within-type variability of 700 hPa winter circulation patterns over the Lake Superior Basin. International Journal of Climatology 19 :41 58. Brinkmann WAR. 1999b. Application of non-hierarchically clustered circulation components to surface weather conditions: Lake Superior Basin winter temperatures.

Theoretical and Applied Climatology 63 :41 56. Brook JR, Samson PJ, Sillman S. 1995. Aggregation of selected three-day periods to estimate annual and seasonal wet deposition totals for sulfate, nitrate, and acidity. Part I: A synoptic and chemical climatology for eastern North America. Journal of Applied Meteorology 34 : 297 325. Bu rger G. 1996. Expanded downscaling for generating local weather scenarios. Climate Research : 111 128. Busuioc A, von Storch H, Schnur R. 1999. Verification of GCM-generated regional season precipitation for current climate and of statistical downscaling estimates

under changing climate conditions. Journal of Climate 12 : 258 272. Cacciamani C, Nanni S, Tibaldi S. 1994. Mesoclimatology of winter temperature and precipitation in the Po Valley of northern Italy. International Journal of Climatology 14 : 777 814. Cavazos T. 1999. Large-scale circulation anomalies conducive to extreme precipitation events and derivation of daily rainfall in northeastern Mexico and southeastern Texas. Journal of Climate 12 : 1506 1523. Caya D, Laprise R. 1999. A semi-implicit semi-Lagrangian regional climate model: the Canadian RCM. Monthly Weather Re iew 127 : 341 362. Caya

D, Laprise R, Giguere M, Bergeron G, Blanchet JP, Stocks BJ, Boer GJ, McFarlane NA. 1995. Description of the Canadian Regional Climate Model. Water Air Soil Pollution 82 : 477 482. Chan JCL, Shi JE. 1997. Application of projection-pursuit principal components analysis method to climate studies. International Journal of Climatology 17 : 103 113. Charles SP, Bates BC, Whetton PH, Hughes JP. 1999. Validation of downscaling models for changed climate conditions: Case study of southwestern Australia. Climate Research 12 :1 14. Chen D, Hellstrom C. 1999. The influence of the North Atlantic

Oscillation on the regional temperature variability in Sweden: spatial and temporal variations. Tellus 51 : 505 516. Christensen JH, Machenhauer B, Jones RG, Scha r C, Ruti PM, Castro M, Visconti G. 1997. Validation of present-day regional climate simulations over Europe: LAM simulations with observed boundary conditions. Climate Dynamics 13 : 489 506. Christensen OB, Christensen JH, Machenhauer B, Botzet M. 1998. Very high-resolution regional climate simulations over Scandinavia present climate. Journal of Climate 11 : 3204 3229. Chu P-S, He Y. 1994. Long-range prediction of Hawaiian winter

rainfall using canonical correlation analysis. International Journal of Climatology 14 : 659 669. Cohen S. 1990. Bringing the global warming issue closer to home: The challenge of regional impact studies. Bulletin of the American Meteorological Society 71 : 520 526. Comrie AC. 1990. The climatology of surface ozone in rural areas: a conceptual model. Progress in Physical Geography 14 : 295 316. Comrie AC. 1992. A procedure for removing the synoptic climate signal from environmental data. International Journal of Climatology 12 : 177 183. Comrie AC. 1994. A synoptic climatology of rural ozone

pollution at three forest sites in Pennsylvania. Atmospheric En ironment 28 : 1601 1614. Copyright  2001 Royal Meteorological Society Int Climatol 21 : 1923 1950 (2001)
Page 22
B. YARNAL ET AL. 1944 Comrie AC. 1996. An all-season synoptic climatology of air pollution in the US Mexico border region. Professional Geographer 48 237 251. Comrie AC, Yarnal B. 1992. Relationships between synoptic-scale atmospheric circulation and ozone concentrations in metropolitan Pittsburgh, Pennsylvania. Atmospheric En ironment 26B : 301 312. Connor GJ, Bonell M. 1998. Air mass and dynamic

parameters affecting trade wind precipitation on the northeast Queensland tropical coast. International Journal of Climatology 18 : 1357 1372. Conway D, Jones PD. 1998. The use of weather types and air flow indices for GCM downscaling. Journal of Hydrology 212 213 348 361. Conway D, Wilby RL, Jones PD. 1996. Precipitation and air flow indices over the British Isles. Climate Research : 169 183. Copeland JH, Pielke RA, Kittel TGF. 1996. Potential climatic impacts of vegetation change: a regional modeling study. Journal of Geophysical Research 101 : 7409 7418. Corte-Real J, Zhang X, Wang X. 1995.

Downscaling GCM information to regional scales: a non-parametric multivariate regression approach. Climate Dynamics 11 : 413 424. Crane RG, Hewitson BC. 1998. Doubled CO2 precipitation changes for the Susquehanna Basin: down-scaling from the GENESIS general circulation model. International Journal of Climatology 18 :65 76. Cress A, Majewski D, Podzun R, Renner V. 1995. Simulation of European climate with a limited area model. Part I: Observed boundary conditions. Contributions to Atmospheric Physics 68 : 161 178. Cui M, von Storch H, Zorita E. 1995. Coastal sea level and the large-scale

climate state. A downscaling exercise for the Japanese Islands. Tellus 47 : 132 144. Arrigo RD, Cook ER, Jacoby GC, Briffa KR. 1993. NAO and sea surface temperature signatures in tree-ring records from the North Atlantic sector. Quaternary Science Re iews 12 : 431 440. Davis RE, Benkovic SR. 1994. Spatial and temporal variations of the January circumpolar vortex over the Northern Hemisphere. International Journal of Climatology 14 : 415 428. Davis RE, Gay DA. 1993a. A synoptic climatological analysis of air quality in the Grand Canyon National Park. Atmospheric En ironment 27A : 713 727. Davis

RE, Gay DA. 1993b. An assessment of air quality variations in the south-western USA using an upper air synoptic climatology. International Journal of Climatology 13 : 755 781. Davis RE, Kalkstein LS. 1990. Development of an automated spatial synoptic climatological classification. International Journal of Climatology 10 : 769 794. Davis RE, Rogers RF. 1992. A synoptic climatology of severe storms in Virginia. Professional Geographer 44 : 319 332. Davis JM, Estis FL, Bloomfield P, Monahan JF. 1991. Complex principal component analysis of sea-level pressure over the eastern USA. International

Journal of Climatology 11 :27 54. Davis RE, Dolan R, Demme G. 1993. Synoptic climatology of Atlantic Coast North-easters. International Journal of Climatology 13 : 171 189. Davis RE, Stanmeyer TM, Jones GV. 1997. A synoptic climatology of tornadoes in Virginia. Physical Geography 18 : 383 407. Deque M, Marquet P, Jones RG. 1998. Simulation of climate change over Europe using a global variable resolution general circulation model. Climate Dynamics 14 : 173 189. Dethloff K, Rinke A, Lehmann R, Christensen JH, Botzet M, Machenhauer B. 1996. Regional climate model of the Arctic atmosphere. Journal

of Geophysical Research 101 : 23401 23422. Diab RD, Preston-Whyte RA, Washington R. 1991. Distribution of rainfall by synoptic type over Natal, South Africa. International Journal of Climatology 11 : 877 888. Dickinson RE, Errico RM, Giorgi F, Bates GT. 1989. A regional climate model for the western United States. Climatic Change 15 383 422. Dilley M. 1996. Synoptic controls on precipitation in the valley of Oaxaca, Mexico. International Journal of Climatology 16 1019 1031. Dilley M. 1997. Climatic factors affecting annual maize yields in the valley of Oaxaca, Mexico. International Journal of

Climatology 17 : 1549 1557. Dorling SR, Davies TD, Pierce CE. 1992a. Cluster analysis: A technique for estimating the synoptic meteorological controls on air and precipitation chemistry method and applications. Atmospheric En ironment 26A : 2575 2581. Dorling SR, Davies TD, Pierce CE. 1992b. Cluster analysis: a technique for estimating the synoptic meteorological controls on air and precipitation chemistry results from Eskdalemuir, south Scotland. Atmospheric En ironment 26A : 2583 2602. Dudek MP, Liang X-Z, Wang W-C. 1996. A regional climate model study of the scale dependence of

cloud-radiation interactions. Journal of Climate : 1221 1234. Easterling D. 1999. Development of regional climate change scenarios using a downscaling approach. Climatic Change 41 : 615 634. El-Kadi AKA, Smithson PA. 1992. Atmospheric classifications and synoptic climatology. Progress in Physical Geography 16 432 455. Enke W, Spekat A. 1997. Downscaling climate model outputs into local and regional weather elements by classification and regression. Climate Research : 195 207. Fernau ME, Samson PJ. 1990. Use of cluster analysis to define periods of similar meteorology and precipitation

chemistry in eastern North America. Part I: Transport patterns. Journal of Applied Meteorology 29 : 735 750. Frakes B, Yarnal B. 1997. A procedure for blending manual and correlation-based synoptic classifications. International Journal of Climatology 17 : 1381 1396. Frey-Buness F, Heimann D, Sauses R. 1995. A statistical-dynamical downscaling procedure for global climate simulations. Theoretical and Applied Climatology 50 : 117 131. Garfin G. 1998. Relationships between winter atmospheric circulation patterns and extreme tree growth anomalies in the Sierra Nevada. International Journal of

Climatology 18 : 725 740. Gatebe CK, Helas G, Tyson PD, Annegarn H, Piketh S. 1999. A seasonal air transport climatology for Kenya. Journal of Geophysical Research 104 : 14237 14244. Copyright  2001 Royal Meteorological Society Int Climatol 21 : 1923 1950 (2001)
Page 23
SYNOPTIC CLIMATOLOGY: DEVELOPMENT AND PROSPECTS 1945 Giorgi F. 1995. Perspectives for regional earth system modeling. Global and Planetary Change 10 :23 42. Giorgi F, Bates GT. 1989. The climatological skill of a regional model over complex terrain. Monthly Weather Re iew 117 2325 2347. Giorgi F, Mearns LO.

1991. Approaches to the simulation of regional climate change: a review. Re iews of Geophysics 29 : 191 216. Giorgi F, Mearns LO. 1999. Introduction to special section: regional climate modeling revisited. Journal of Geophysical Research 104 : 6335 6352. Giorgi F, Marinucci MR, Visconti G. 1992. A 2 CO climate change scenario over Europe generated using a limited area model nested in a general circulation model. Part 2: Climate change scenario. Journal of Geophysical Research 97 : 10011 10028. Giorgi F, Marinucci MR, Bates GT. 1993a. Development of a second-generation regional climate model

(RegCM2). Part I: Boundary-layer and radiative transfer processes. Monthly Weather Re iew 121 : 2794 2813. Giorgi F, Marinucci MR, Bates GT, DeCanio G. 1993b. Development of a second-generation regional climate model (RegCM2). Part II: Convective processes and assimilation of lateral boundary conditions. Monthly Weather Re iew 121 : 2794 2813. Giorgi F, Bates GT, Nieman SJ. 1993c. The multiyear surface climatology of a regional atmospheric model over the western United States. Journal of Climate :75 95. Giorgi F, Brodeur CS, Bates GT. 1994a. Regional climate change scenarios over the United

States produced with a nested regional climate model. Journal of Climate : 375 399. Giorgi F, Hostetler SW, Brodeur CS. 1994b. Analysis of the surface hydrology in a regional climate model. Quarterly Journal of the Royal Meteorological Society 120 : 161 183. Giorgi F, Mearns LO, Shields C, Mayer L. 1996. A regional model study of the importance of local versus remote controls of the 1988 drought and 1993 flood over the central United States. Journal of Climate : 1150 1162. Goodess CM, Palutikof JP. 1998. Development of daily rainfall scenarios for southeast Spain using a circulation-type

approach to downscaling. International Journal of Climatology 10 : 1051 1083. Goyette S, Laprise R. 1996. Numerical investigation with a physically based regional interpolator for off-line downscaling of GCMs: FIZR. Journal of Climate : 3464 3495. Greene JS. 1996. A synoptic climatological analysis of summertime precipitation intensity in the eastern United States. Physical Geography 17 : 401 418. Greene JS, Kalkstein LS, Ye H, Smoyer K. 1999. Relationships between synoptic climatology and atmospheric pollution at four US cities. Theoretical and Applied Climatology 62 : 163 174. Grundstein AJ,

Leathers DJ. 1999. A spatial analysis of snow-surface energy exchanges over the northern Great Plains of the United States in relation to synoptic scale forcing mechanisms. International Journal of Climatology 19 : 489 511. Gutowski WJ Jr, Wilby RL, Hay LE, Anderson CJ, Arritt RW, Clark MP, Leavesley GH, Pan Z, da Silva R, Takle ES. 2000. Statistical and dynamical downscaling of global model output for US National Assessment hydrological analyses. 11 th Symposium on Global Change Studies Long Beach CA 14 January 2000 . http: // www.pircs.iastate.edu abstracts STATLAMAMS.pdf Gyalistras D, von

Storch H, Fischlin A, Beniston M. 1994. Linking GCM-simulated climatic changes to ecosystem models: Case studies of statistical downscaling in the Alps. Climate Research : 167 189. Harnack RP, Jensen DT, Cermak III JR. 1998. Investigation of upper-air conditions occurring with heavy summer rain in Utah. International Journal of Climatology 18 : 701 723. Hartley S, Keables MJ. 1998. Synoptic associations of winter climate and snowfall variability in New England, USA, 1950 1992. International Journal of Climatology 18 : 281 298. Harvey VL, Hitchman MH. 1996. A climatology of the Aleutian High.

Journal of the Atmospheric Sciences 53 : 2088 2101. Hewitson BC, Crane RG. 1994. Precipitation controls in Southern Mexico. In Neural Nets Applications in Geography , Hewitson BC, Crane RG (eds). Kluwer Academic Press: Dordrecht; 121 142. Hewitson BC, Crane RG. 1996. Climate downscaling: Techniques and application. Climate Research :85 95. Heyen H, Zorita E, von Storch H. 1996. Statistical downscaling of monthly mean north Atlantic air-pressure to sea level anomalies in the Baltic Sea. Tellus 48 : 312 323. Hirakuchi H, Giorgi F. 1995. Multiyear present-day and 2 CO simulations of monsoon

climate over eastern Asia and Japan with a regional climate model nested in a general circulation model. Journal of Geophysical Research 100 : 21105 21125. Hirschboeck KK, Ni F, Wood ML, Woodhouse CA. 1996. Synoptic dendroclimatology: overview and prospectus. In Tree Rings En ironment and Humanity Proceedings of the International Conference . Tucson, Arizona, 17 21 May 1994, Dean JS, Meko DM, Swetnam TW (eds). The University of Arizona: Tuscon, AZ; 205 223. Holt T. 1999. A classification of ambient climate conditions during extreme surge events off western Europe. International Journal of

Climatology 19 : 725 744. Hong S-Y, Leetmaa A. 1999. An evaluation of the NCEP RSM for regional climate modeling. Journal of Climate 12 : 592 609. Hostetler SW, Bates GT, Giorgi F. 1993. Interactive coupling of a lake thermal model with a regional climate model. Journal of Geophysical Research 98 : 5045 5057. Hostetler SW, Giorgi F. 1995. Effects of a 2 CO climate on two large lake systems: Pyramid Lake, Nevada, and Yellowstone Lake, Wyoming. Global and Planetary Change 10 :43 54. Hughes JP, Lettenmaier DP, Guttorp P. 1993. A stochastic approach for assessing the effect of changes in synoptic

circulation patterns on gauge precipitation. Water Resources Research 29 : 3303 3315. Huth R. 1997. Potential of continental-scale circulation for determination of local daily surface variables. Theoretical and Applied Climatology 56 : 165 186. Inamdar SR, Singh SV. 1993. Relationship between circulation and rainfall over India during the Southwest Monsoon Season. Part II: 700 hPa contour height field and some synoptic indices. Theoretical and Applied Climatology 48 :1 14. Jacob D, Podzun R. 1997. Sensitivity studies with the regional climate model REMO. Meteorology and Atmosperic Physics 63

119 129. Jones PD, Hulme M, Briffa KR. 1993. A comparison of Lamb circulation types with an objective classification scheme. International Journal of Climatology 13 : 655 663. Copyright  2001 Royal Meteorological Society Int Climatol 21 : 1923 1950 (2001)
Page 24
B. YARNAL ET AL. 1946 Jones RG, Murphy JM, Noguer M. 1995. Simulation of climate change over Europe using a nested regional-climate model. I: Assessment of control climate, including sensitivity to location of lateral boundaries. Quarterly Journal of the Royal Meteorological Society 121 : 1413 1449. Jones RG, Murphy

JM, Noguer M, Keen AB. 1997. Simulation of climate change over Europe using a nested regional-climate model. Part II: Comparison of driving and regional model responses to a doubling of carbon dioxide. Quarterly Journal of the Royal Meteorological Society 123 : 265 292. Joubert AM, Hewitson BC. 1997. Simulating present and future climates of southern Africa using general circulation models. Progress in Physical Geography 21 :51 78. Juang H-MH, Kanamitsu M. 1994. The NMC nested regional spectral model. Monthly Weather Re iew 122 :3 26. Juang H-M, Hong S-Y, Kanamitsu M. 1997. The NCEP regional

spectral model: an update. Bulletin of the American Meteorological Society 78 : 2125 2143. Kalkstein LS, Nichols MC, Barthel CD, Greene JS. 1996. A new spatial synoptic classification: application to air-mass analysis. International Journal of Climatology 16 : 938 1004. Kalkstein LS, Sheridan SC, Graybeal DY. 1998. A determination of character and frequency changes in air masses using a spatial synoptic classification. International Journal of Climatology 18 : 1223 1236. Kalnay E, Kanamitsu M, Kistler R, Collins W, Deaven D, Gandin L, Iredell M, Saha S, White G, Woollen J, Zhu Y, Chelliah M,

Ebisuzaki W, Higgins W, Janowiak J, Mo KC, Ropelewski C, Wang J, Leetma A, Reynolds R, Jenne R, Joseph D. 1996. The NCEP NCAR 40-year reanalysis project. Bulletin of the American Meteorological Society 77 : 437 471. Kang I-S, Ho C-H, Lim Y-K, Lau K-M. 1999. Principal modes of climatological seasonal and intraseasonal variations of the Asian Summer Monsoon. Monthly Weather Re iew 127 : 322 340. Karl TR, Wang W-C, Schlesinger ME, Knight RW, Portman D. 1990. A method of relating general circulation model simulated climate to the observed local climate. Part I: Seasonal statistics. Journal of

Climate : 1053 1079. Karl TR, Knight RW, Easterling DR, Quayle RG. 1996. Indices of climate change for the United States. Bulletin of the American Meteorological Society 77 : 279 292. Kattenberg A, Giorgi F, Grassl H, Meehl GA, Mitchell JFB, Stouffer RJ, Tokioka T, Weaver AJ, Wigley TML. 1996. Climate models projections of future climate. In Climate Change 1995: The Science of Climate Change , Houghton JT, Meira Filho LG, Callander BA, Harris N, Kattenberg A, Maskell K (eds). Cambridge University Press: Cambridge; 285 357. Katz RW, Parlange MB. 1996. Mixtures of stochastic processes:

application to statistical downscaling. Climate Research : 185 193. Katzfey JJ. 1995. Simulation of extreme New Zealand precipitation events. I: Sensitivity to orography and resolution. Monthly Weather Re iew 123 : 737 754. Kaufmann RK, Snell SE, Gopal S, Dezzani R. 1999. The significance of synoptic patterns identified by the Kirchhofer Technique: a Monte Carlo approach. International Journal of Climatology 19 : 619 626. Kida H, Koide T, Sasaki H, Chiba M. 1991. A new approach for coupling a limited area model to a GCM. Journal of the Meteorological Society of Japan 69 : 723 727. Kidson JW.

1994a. An automated procedure for the identification of synoptic types applied to the New Zealand region. International Journal of Climatology 14 : 711 721. Kidson JW. 1994b. Relationship of New Zealand daily and monthly weather patterns to synoptic weather types. International Journal of Climatology 14 : 723 737. Kidson JW. 1997. The utility of surface and upper air data in synoptic climatological specification of surface climatic variables. International Journal of Climatology 17 : 399 413. Kidson JW, Thompson CS. 1998. A comparison of statistical and model-based downscaling techniques for

estimating local climate variations. Journal of Climate 11 : 735 753. Kidson JW, Watterson IG. 1995. A synoptic climatological evaluation of the changes in the CSIRO nine-level model with doubled CO in the New Zealand region. International Journal of Climatology 15 : 1179 1194. Kim J. 1997. Precipitation and snow budget over the southwestern United States during the 1994 1995 winter season in a mesoscale model simulation. Water Resources Research 33 : 2831 2839. Kim J, Miller NL, Georgakakos KP, Guetter AK. 1998. River flow response to precipitation and snow budget in California during the

1994 1995 winter. Journal of Climate 11 : 2378 2386. Knappenberger PC, Michaels PJ. 1993. Cyclone tracks and wintertime climate in the mid-Atlantic region of the USA. International Journal of Climatology 13 : 509 531. Kohonen T. 2001. Self Organizing Maps (3rd edn). Springer: Berlin. Konrad CE. 1997. Synoptic-scale features associated with warm season heavy rainfall over the interior southeastern United States. Weather and Forecasting 12 : 557 571. Konrad CE. 1998. Persistent planetary scale circulation patterns and their relationship with cold air outbreak activity over the eastern United

States. International Journal of Climatology 18 : 1209 1221. Kripalani RH, Kulkarni A. 1998. The relationship between some large-scale atmospheric patterns and rainfall over Southeast Asia: a comparison with features over India. Theoretical and Applied Climatology 59 :1 11. Kutiel H, Maheras P, Guika S. 1996. Circulation indices and their relationship with rainfall conditions across the Mediterranean. Theoretical and Applied Climatology 54 : 125 138. Lakhtakia MN, Yarnal B, Johnson DL, White RA, Miller DA, Yu Z. 1998. A simulation of river-basin response to mesoscale meteorological forcing:

the Susquehanna River Basin Experiment (SRBEX). Journal of the American Water Resources Association 34 : 921 937. Lakhtakia MN, Yu Z, Yarnal B, White RA, Miller DA. 1999. Sensitivity of simulated surface runoff to mesoscale meteorological model resolution in a linked-model experiment. Climate Research 12 :15 27. Lana X, Fernandez Mills G. 1994. Minimum sample size for synoptic weather type classification. Application to winter period data recorded on the Catalan Coast (northeast Spain). International Journal of Climatology 14 : 1051 1060. Langmann B, Graf H-F. 1997. The chemistry of the

polluted atmosphere over Europe: simulations and sensitivity studies with a regional chemistry-transport-model. Atmospheric En ironment 31 : 3239 3257. Copyright  2001 Royal Meteorological Society Int Climatol 21 : 1923 1950 (2001)
Page 25
SYNOPTIC CLIMATOLOGY: DEVELOPMENT AND PROSPECTS 1947 Langmann B, Herzog M, Graf H-F. 1998. Radiative forcing of climate by sulfate aerosols as determined by a regional circulation chemistry transport model. Atmospheric En ironment 32 : 2757 2768. Laprise R, Caya D, Giguere M, Bergeron G, Cote H, Blanchet J-P, Boer GJ, McFarlane NA. 1998.

Climate and climate change in western Canada as simulated by the Canadian regional climate model. Atmosphere Ocean 36 : 119 167. Lau K-H, Wang A-Y, Kuo Y-H, Chen S-J, Dudhia J. 1998. The evolution of the East Asia summer monsoon in June 1994: Numerical simulations. Journal of the Meteorological Society of Japan 76 : 749 764. Leathers DJ. 1993. A synoptic climatology of northeastern United States tornadoes. Physical Geography 14 : 171 190. Leathers DJ, Ellis AW. 1996. Synoptic mechanisms associated with snowfall increases to the lee of Lakes Erie and Ontario. International Journal of

Climatology 16 : 1117 1135. Lenters JD, Cook KH. 1999. Summertime precipitation variability over South America: role of the large-scale circulation. Monthly Weather Re iew 127 : 409 431. Leung LR, Ghan SJ. 1995. A subgrid parameterization of orographic precipitation. Theoretical and Applied Climatology 52 :95 118. Leung LR, Ghan SJ. 1998. Parameterizing subgrid orographic precipitation and surface cover in climate models. Monthly Weather Re iew 126 : 3271 3291. Leung LR, Ghan SJ. 1999a. Pacific Northwest climate sensitivity simulated by a regional climate model driven by a GCM. Part I: Control

simulations. Journal of Climate 12 : 2010 2030. Leung LR, Ghan SJ. 1999b. Pacific Northwest climate sensitivity simulated by a regional climate model driven by a GCM. Part II: CO simulations. Journal of Climate 12 : 2031 2053. Leung LR, Wigmosta MS. 1999. Potential climate change impacts on mountain watersheds in the Pacific Northwest. Journal of the American Water Resources Association 35 : 1463 1471. Leung LR, Wigmosta MS, Ghan SJ, Epstein DJ, Vail LW. 1996. Application of a subgrid orographic precipitation surface hydrology scheme to a mountain watershed. Journal of Geophysical Research 101

: 12803 12817. Leung LR, Hamlet AF, Lettenmaier DP, Kumar A. 1999a. Simulations of the ENSO hydroclimate signals in the Pacific Northwest Columbia River basin. Bulletin of the American Meteorological Society 80 : 2313 2329. Leung LR, Ghan SJ, Zhao Z-C, Luo Y, Wang W-C, Wei H-L. 1999b. Intercomparison of regional climate simulations of the 1991 summer monsoon over eastern Asia. Journal of Geophysical Research 104 : 12803 12817. Levey KM. 1996. Interannual temperature variability and associated synoptic climatology at Cape Town. International Journal of Climatology 16 : 293 306. Lorenz EN. 1956.

Empirical orthogonal functions and statistical weather prediction. Scientific Report No. 1, Contract AF19 (604)-1566, Meteorology Department, Massachusetts Institute of Technology, Cambridge, MA. Liu Y, Giorgi F, Washington WM. 1994. Simulation of summer monsoon climate over East Asia with an NCAR regional climate model. Monthly Weather Re iew 122 : 2331 2349. Liu Y, Avissar R, Giorgi F. 1996. Simulation with the regional climate model RegCM2 of extremely anomalous precipitation during the 1991 east Asian flood: an evaluation study. Journal of Geophysical Research 101 : 26199 26215. Lu thi D,

Cress A, Davies HC, Frei C, Scha r C. 1996. Interannual variability and regional climate simulations. Theoretical and Applied Climatology 53 : 185 209. Lynch AH, Chapman WL, Walsh JE, Weller G. 1995. Development of a regional climate model of the western Arctic. Journal of Climate : 1555 1570. Lynch AH, Glueck MF, Chapman WL, Bailey DA, Walsh JE. 1997. Satellite observation and climate system model simulation of the St. Lawrence Island polynya. Tellus 49A : 277 297. Lynch AH, McGinnis DL, Bailey DA. 1998. Snow-albedo feedback and the spring transition in a regional climate system model:

influence of land surface model. Journal of Geophysical Research 103 : 29037 29049. Lynch AH, Bonan GB, Chapin FS III, Wu W. 1999a. Impact of tundra ecosystems on the surface energy budget and climate of Alaska. Journal of Geophysical Research 104 : 6647 6660. Lynch AH, Chapin FS III, Hinzman LD, Wu W, Lilly E, Vourlitis G, Kim E. 1999b. Surface energy balance on the Arctic tundra: measurements and models. Journal of Climate 12 : 2585 2606. Lyons WF, Bonell M. 1992. Daily meso-scale rainfall in the tropical wet dry climate of the Townsville Area, north-east Queensland during the 1998 1989 wet

season: Synoptic-scale airflow consideration. International Journal of Climatology 12 : 655 684. Lyons WF, Bonell M. 1994. Regionalization of daily mesoscale rainfall in the tropical wet dry climate of the Townsville Area of North-east Queensland during the 1988 1989 wet season. International Journal of Climatology 14 : 135 163. Maak K, von Storch H. 1997. Statistical downscaling of monthly mean air temperature to the beginning of flowering of Galanthus ni alis L. in northern Germany. Journal of Biometeorology 41 :5 12. MacKay MD, Stewart RE, Bergeron G. 1998. Downscaling the hydrological

cycle in the Mackenzie Basin with the Canadian Regional Climate Model. Atmosphere Ocean 36 : 179 211. Marinucci MR, Giorgi F, Beniston M, Wild M, Tschuck P, Ohmura A, Bernasconi A. 1995. High resolution simulations of January and July climate over the western Alpine region with a nested regional modeling system. Theoretical and Applied Climatology 51 119 138. Marroquin A, Garcia JA, Garrido J, Mateos VL. 1995. Neyman-Scott Cluster Model for daily rainfall processes in Lower Extremadura (Spain): rainfall generating mechanisms. Theoretical and Applied Climatology 52 : 183 193. Marshall GJ,

Turner J, Miners WD. 1998. Interpreting recent accumulation records through an understanding of the regional synoptic climatology: an example from the southern Antarctic Penninsula. Annals of Glaciology 27 : 610 616. Martin E, Timbal B, Brun E. 1997. Downscaling of general circulation model outputs: Simulation of the snow climatology of the French Alps and sensitivity to climate change. Climate Dynamics 13 :45 56. Matyasovszky I, Boga rdi I. 1996. Downscaling two versions of a general circulation model (GCM) to estimate local hydroclimatic factors under climate change. Hydrological Sciences

Journal 41 : 117 129. McGinnis D. 1994. Predicting snowfall from synoptic circulation: a comparison of linear regression and neural network methodologies. In Neural Nets Applications in Geography , Hewitson BC, Crane RG (eds). Kluwer Academic Publishers: Dordrecht; 79 94. Copyright  2001 Royal Meteorological Society Int Climatol 21 : 1923 1950 (2001)
Page 26
B. YARNAL ET AL. 1948 McGregor JL. 1997. Regional climate modelling. Meteorology and Atmospheric Physics 63 : 105 117. McGregor GR, Bamzelis D. 1995. Synoptic typing and its application to the investigation of weather air

pollution relationships, Birmingham, United Kingdom. Theoretical and Applied Climatology 51 : 223 236. McGregor JL, Walsh KJ. 1993. Nested simulations of perpetual January climate over the Australian region. Journal of Geophysical Research 98 : 23283 23290. McGregor JL, Walsh KJ. 1994. Climate change simulations of Tasmanian precipitation using multiple nesting. Journal of Geophysical Research 99 : 20889 20905. Mearns LO, Rosenzweig C, Goldberg R. 1997. Mean and variance change in climate scenarios: methods, agricultural applications, and measures of uncertainty. Climatic Change 35 : 367 396.

Mearns LO, Boga rdi I, Matyasovszky I, Palecki M. 1999a. Comparison of climate change scenarios generated from regional climate model experiments and statistical downscaling. Journal of Geophysical Research 104 : 6603 6621. Mearns LO, Mavromatis T, Tsvetsinskaya E, Hays C, Easterling W. 1999b. Comparative responses of EPIC and CERES crop models to high and low spatial resolution climate change scenarios. Journal of Geophysical Research 104 : 6603 6621. Miller NL, Kim J. 1996. Numerical prediction of precipitation and river flow over the Russian River watershed during the January 1995

California storms. Bulletin of the American Meteorological Society 77 : 101 105. Miller NL, Kim J, Hartman RK, Farrara J. 1999. Downscaled climate and streamflow study of the southwestern United States. Journal of the American Water Resources Association 35 : 1525 1537. Mock CJ, Bartlein PJ, Anderson PM. 1998. Atmospheric circulation patterns and spatial climatic variations in Beringia. International Journal of Climatology 18 : 1085 1104. Moron V. 1997. Trend, decadal, and interannual variability in annual rainfall of subequatorial and tropical North Africa (1990 1994). International Journal

of Climatology 17 : 785 805. Mote TL. 1998a. Mid-tropospheric circulation and surface melt on the Greenland Ice Sheet. Part I: Atmospheric teleconnections. International Journal of Climatology 18 : 111 129. Mote TL. 1998b. Mid-tropospheric circulation and surface melt on the Greenland Ice Sheet. Part II: Synoptic climatology. International Journal of Climatology 18 : 131 145. Murphy J. 1999. An evaluation of statistical and dynamical techniques for downscaling local climate. Journal of Climate 12 2256 2284. Neale SM, Fitzharris BB. 1997. Energy balance and synoptic climatology of a melting

snowpack in the southern Alps, New Zealand. International Journal of Climatology 17 : 1595 1609. Noguer M, Jones R, Murphy J. 1998. Sources of systematic errors in the climatology of a regional climate model over Europe. Climate Dynamics 14 : 691 712. Pan Z, Segal M, Arritt RW, Chen T-C, Weng S-P. 1999a. A method for simulating effects of quasi-stationary wave anomalies on regional climate. Journal of Climate 12 : 1336 1343. Pan Z, Takle ES, Segal M, Arritt RW. 1999b. Simulation of potential impacts of man-made land use changes on US summer climate under various synoptic regimes. Journal of

Geophysical Research 104 : 6515 6528. Pan Z, Christensen JH, Arritt RW Jr, Gutowski WJ, Takle ES. 2000. Ten-year US regional climate simulations for impacts assessments. 11 th Symposium on Global Change Studies Long Beach CA 14 January 2000 . http: // www.pircs. iastate.edu abstracts ams00 LB 10yr.pdf Palutikof J, Winkler J, Goodess C, Andresen J. 1997. The simulation of daily temperature time series from GCM output. Part I: Comparison of model data with observations. Journal of Climate 10 : 2497 2513. Pielke RA, Cotton WR, Walko RL, Tremback CJ, Lyons WA, Grasso LD, Nicholls ME, Moran MD,

Wesley DA, Lee TJ, Copeland JH. 1992. A comprehensive meteorological modeling system RAMS. Meteorology and Atmospheric Physics 49 :69 91. Podzun R, Cress A, Majewski D, Renner V. 1995. Simulation of European climate with a limited area model. Part I: Observed boundary conditions. Contributions to Atmospheric Physics 68 : 205 225. Prezerakos NG. 1998. Lower tropospheric structure and synoptic scale circulation patterns during prolonged temperature inversions over Athens, Greece. Theoretical and Applied Climatology 60 :63 76. Renwick JA, Katzfey JJ, Nguyen KC, McGregor JL. 1998. Regional model

simulations of New Zealand climate. Journal of Geophysical Research 103 : 5973 5982. Richardson C. 1981. Stochastic simulation of daily precipitation, temperature, and solar radiation. Water Resources Research 17 182 190. Rinke A, Dethloff K. 2000. On the sensitivity of a regional Arctic climate model to initial and boundary conditions. Climate Research 14 : 101 113. Rinke A, Dethloff K, Christensen JH, Botzet M, Machenhauer B. 1997. Simulation and validation of Arctic radiation and clouds in a regional climate model. Journal of Geophysical Research 102 : 29833 29847. Robock A, Turco R,

Harwell M, Ackerman T, Andressen R, Chang H-S, Sivakumar M. 1993. Use of general circulation model (GCM) output in the creation of climate change scenarios for impact analysis. Climatic Change 23 : 293 335. Rodriguez-Puebla C, Encinas AH, Nieto S, Garmendia J. 1998. Spatial and temporal patterns of annual precipitation variability over the Iberian Peninsula. International Journal of Climatology 18 : 299 316. Romero R, Sumner G, Ramis C, Genoves A. 1999. A classification of the atmospheric circulation patterns producing significant daily rainfall in the spanish Mediterranean area. International

Journal of Climatology 19 : 765 785. Roswintiarti O, Niyogi DS, Raman S. 1998. Teleconnections between tropical Pacific sea surface temperature anomalies and North Carolina precipitation anomalies during El Nin o events. Geophysical Research Letters 25 : 4201 4204. Ruiz NE, Vargas WM. 1998. 500 hPa vorticity analyses over Argentina: their climatology and capacity to distinguish synoptic-scale precipitation. Theoretical and Applied Climatology 60 :77 92. Rutlant J, Fuenzalida H. 1991. Synoptic aspects of the central Chile rainfall variability associated with the Southern Oscillation.

International Journal of Climatology 11 :63 76. Sailor D, Li X. 1999. A semiempirical downscaling approach for predicting regional temperature impacts associated with climate change. Journal of Climate 12 : 103 114. Copyright  2001 Royal Meteorological Society Int Climatol 21 : 1923 1950 (2001)
Page 27
SYNOPTIC CLIMATOLOGY: DEVELOPMENT AND PROSPECTS 1949 Saunders I, Byrne J. 1999. Using synoptic surface and geopotential height fields for generating grid-scale precipitation. International Journal of Climatology 19 : 1165 1176. Scha rC,FreiC,Lu thi D, Davies HC. 1996. Surrogate

climate-change scenarios for regional climate models. Geophysical Research Letters 23 : 669 672. Scha rC,Lu thi D, Beyerle U, Heise E. 1999. The soil-precipitation feedback: a process study with a regional climate model. Journal of Climate 12 : 722 741. Schnur R, Lettenmaier D. 1998. A case study of statistical downscaling in Australia using weather classification by recursive partitioning. Journal of Hydrology 212 213 : 362 379. Schubert S. 1998. Downscaling local extreme temperature changes in south-eastern Australia from CSIRO Mark 2 GCM. International Journal of Climatology 18 : 1419 1438.

Schubert S, Henderson-Sellers A. 1997. A statistical model to downscale local daily temperature extremes from synoptic-scale circulation patterns in the Australian region. Climate Dynamics 13 : 223 234. Schubert SD, Helfand HM, Wu C-Y, Min W. 1998. Subseasonal variations in warm-season moisture transport and precipitation over the central and eastern United States. Journal of Climate 11 : 2530 2555. Semenov M, Barrow E. 1997. Use of a stochastic weather generator in the development of climate change scenarios. Climatic Change 35 : 397 414. Serra C, Fernandez Mills G, Periago MC, Lana X. 1998.

Surface synoptic circulation and daily precipitation in Catalonia. Theoretical and Applied Climatology 59 :29 49. Seth A, Giorgi F. 1996. Three-dimensional model study of organized mesoscale circulations induced by vegetation. Journal of Geophysical Research 101 : 7371 7391. Shahgedanova M, Burt TP, Davies TD. 1998. Synoptic climatology of air pollution in Moscow. Theoretical and Applied Climatology 61 :85 102. Singh CV. 1999. Principal components of monsoon rainfall in normal, flood and drought years over India. International Journal of Climatology 19 : 639 652. Small EE, Giorgi F, Sloan LC.

1999. Regional climate model simulation of precipitation in central Asia: Mean and interannual variability. Journal of Geophysical Research 104 : 6563 6582. Solman S, Nun ez M. 1999. Local estimates of global climate change: A statistical downscaling approach. International Journal of Climatology 19 : 835 861. Speer MS, Leslie LM. 1997. A climatology of coastal ridging over south-eastern Australia. International Journal of Climatology 17 831 845. Spellman G. 1997. Evaluating indices of mid-latitude circulation. Journal of Meteorology 22 : 237 242. Stefanicki G, Talkner P, Weber RO. 1998.

Frequency changes of weather types in the Alpine region since 1945. Theoretical and Applied Climatology 60 :47 61. Stonefelt MD, Fontaine TA, Hotchkiss RH. 2000. Impacts of climate change on water yield in the Upper Wind River Basin. Journal of the American Water Resources Association 36 : 321 336. Sumner G. 1996. Daily precipitation patterns over Wales: towards a detailed precipitation climatology. Transactions Institute of British Geographers 21 : 157 176. Sumner G, Guijarro JA, Ramis C. 1995. The impact of surface circulation on significant daily rainfall patterns over Mallorca.

International Journal of Climatology 15 : 673 696. Svensson C. 1999. Empirical orthogonal function analysis of daily rainfall in the upper Reaches of the Huai River Basin, China. Theoretical and Applied Climatology 62 : 147 161. Sweeney JC, O Hare GP. 1992. Geographical variations in precipitation yields and circulation types in Britain and Ireland. Transactions Institute of British Geographers 17 : 448 463. Takle ES, Gutowski WJ Jr, Arritt RW, Pan Z, Anderson CJ, da Silva R, Caya D, Chen S-C, Giorgi F, Christensen JH, Hong S-Y, Juang H-MH, Katzfey JJ, Lapenta WM, Laprise R, Liston G, Lopez P,

Pielke RA Sr, McGregor J, Roads JO. 1999. Project to Intercompare Regional Climate Simulations (PIRCS): Description and initial results. Journal of Geophysical Research 104 19443 19461. van Lipzig NPM, van Meijgaard E, Oerlemans J. 1998. Evaluation of a regional atmospheric model for January 1993, using in situ measurements from the Antarctic. Annals of Glaciology 27 : 507 514. van Lipzig NPM, van Meijgaard E, Oerlemans J. 1999. Evaluation of a regional atmospheric model using measurements of surface heat exchange processes from a site in Antarctica. Monthly Weather Re iew 127 : 1994 2011. van

Regenmortel G. 1995. Regionalization of Botswana rainfall during the 1980s using principal components analysis. International Journal of Climatology 15 : 313 323. Venugopal V, Foufoula-Georgiou E, Sapozhnikov V. 1999. A space-time downscaling model for rainfall. Journal of Geophysical Research 104 : 19705 19721. von Storch H, Zorita E, Cubasch U. 1993. Downscaling of global climate change estimates to regional scales: An application to Iberian rainfall in wintertime. Journal of Climate : 1161 1171. Walland DJ, Simmonds I. 1997. Association between modes of variability of January Northern

Hemisphere snow cover and circulation. Theoretical and Applied Climatology 58 : 197 210. Walsh K, McGregor J. 1995. January and July climate simulations over the Australian region using a limited-area model. Journal of Climate : 2387 2403. Walsh K, McGregor J. 1996. Simulations of Antarctic climate using a limited area model. Journal of Geophysical Research 101 19093 19108. Walsh K, McGregor J. 1997. An assessment of simulations of climate variability over Australia with a limited area model. International Journal of Climatology 17 : 201 223. Weber RO. 1998. Climatology of regional flow

patterns around Basel. Theoretical and Applied Climatology 59 :13 27. Weichert A, Bu rger G. 1998. Linear versus nonlinear techniques in downscaling. Climate Research 10 :83 93. Copyright  2001 Royal Meteorological Society Int Climatol 21 : 1923 1950 (2001)
Page 28
B. YARNAL ET AL. 1950 White WB, Cherry NJ. 1999. Influence of the antarctic circumpolar wave upon New Zealand temperature and precipitation during autumn-winter. Journal of Climate 12 : 960 976. Wilby RL. 1994. Stochastic weather type simulation for regional climate change impact assessment. Water Resources

Research 30 3395 3403. Wilby RL. 1995. Greenhouse hydrology. Progress in Physical Geography 19 : 351 369. Wilby RL. 1997. Non-stationarity in daily precipitation series: Implications for GCM down-scaling using atmospheric circulation indices. International Journal of Climatology 17 : 439 454. Wilby RL, Wigley TML. 1997. Downscaling general circulation model output: a review of methods and limitations. Progress in Physical Geography 21 : 530 548. Wilby RL, Wigley TML, Conway D, Jones PD, Hewitson BC, Main J, Wilks DS. 1998a. Statistical downscaling of general circulation model output: a

comparison of methods. Water Resources Research 34 : 2995 3008. Wilby RL, Hassan H, Hanaki K. 1998b. Statistical downscaling of hydrometeorological variables using general circulation model output. Journal of Hydrology 205 :1 19. Wilks DS. 1999. Multisite downscaling of daily precipitation with a stochastic weather generator. Climate Research 11 : 125 136. Winkler JA, Palutikof JP, Andresen JA, Goodess C. 1997. The simulation of daily temperature time series from GCM output. Part II: Sensitivity analysis of an empirical transfer function methodology. Journal of Climate 10 : 2514 2532.

Woodhouse CA. 1997. Winter climate and atmospheric circulation patterns in the Sonoran Desert region, USA. International Journal of Climatology 17 : 859 873. Xu C-Y. 1999. From GCMs to river flow: a review of downscaling methods and hydrologic modeling approaches. Progress in Physical Geography 23 : 229 249. Yao CS. 1998. A loading correlation model for climatic classification in terms of synoptic climatology. Theoretical and Applied Climatology 61 : 113 120. Yarnal B. 1993. Synoptic Climatology in En ironmental Analysis . Belhaven Press: London. Yarnal B, Draves JD. 1993. A synoptic

climatology of stream flow and acidity. Climate Research : 193 202. Yarnal B, Frakes B. 1997. Using synoptic climatology to define representative discharge events. International Journal of Climatology 17 : 323 341. Yarnal B, Lakhtakia MN, Yu Z, White RA, Pollard D, Miller DA, Lapenta WM. 2000. A Linked meteorological and hydrological model system: The Susquehanna River Basin Experiment (SRBEX). Global and Planetary Change 25 : 149 161. Yu Z, Lakhtakia MN, Yarnal B, White RA, Miller DA, Frakes B, Barron EJ, Duffy C, Schwartz FW. 1999. Simulating the river-basin response to atmospheric forcing

by linking a mesoscale meteorological model and a hydrologic model system. Journal of Hydrology 218 :72 91. Zelenka MP. 1997. An analysis of the meteorological parameters affecting ambient concentrations of acid aerosols in Uniontown, Pennsylvania. Atmospheric En ironment 31 : 869 878. Zorita E, von Storch H. 1999. The analog method as a simple statistical downscaling technique: Comparison with more complicated methods. Journal of Climate 12 : 2474 2489. Copyright  2001 Royal Meteorological Society Int Climatol 21 : 1923 1950 (2001)