An enhanced hydro-ecological model (RHESSys) to explore climate change interactions between - PowerPoint Presentation

An enhanced hydro-ecological model (RHESSys) to explore climate change interactions between precipitation patterns, topography and forests in a New York City water supply watershed Antoine L. Randolph1,3, Lawrence E. Band1, Christina L. Tague2, Matthew B. Dickinson4, Elliot M. Schneiderman5 1 University of North Carolina Chapel Hill, Department of Geography 2 University of California Santa Barbara, Bren School of Environmental Science3 CUNY Institute for Sustainable Cities, Hunter College4 U.S Forest Service Northeast Research Station, Delaware OH5 New York City Environmental Protection, Mapping and Modeling Watershed/Tifft Science and Technical Symposium, 18-19 September 2013, West Point New York

Presentation Outline Hydro-ecological models as management toolsA brief overview of RHESSysEnhancements to the baseline version of RHESSysImplementation of RHESSys for Biscuit BrookExamples of Biscuit Brook model outputFuture research and development

Hydro-ecological models as management tools tools3 Can be used to forecast the potential impacts of climate change on forest structure and composition:changes in the frequency of extreme weather eventswind damage, ice damage, floodingcanopy damagechange in precipitation patternsincreased stress greater susceptibility to insects and pathogensdecrease in the viability of seedchanges in the distribution of tree species habitat lossloss of commercially important tree speciesChanges in forest cover can affect the quantity and quality of stream flowincreased erosion, turbidity and nutrient loadingchanges in the spatial and temporal availability of water

A brief overview of RHESSys4

RHESSys genealogy5(RHESSys)

RHESSys SpecificsRegional Hydro-Ecological Simulation SystemGIS-based modelSpatially distributedHierarchical/NestedWatershed modelsmall scale (e.g., a 10,000m2 catchment, 1st order streams)regional scale (e.g., a 1000km2 basin, 4th order streams)minimum pixel (i.e., landscape patch) size 5 to 10m2 Process modelecological processes (e.g., photosynthesis)hydrological processes (e.g., infiltration, run-off)explicit routing (fully distributed – extended from DHSVM routing)Topmodel (quasi-distributed based on wetness index)6

Model Enhancements7

Vegetation modeling8Tree dimensions define zones of influence for each species, within which the characteristics of the tree modify local microclimate. Canopy dominant trees function as “Keystone species.”

RHESSys modeling enhancements (highlights)explicit modeling of tree speciesleaf C/N ratio, specific leaf area, environmental tolerances, dynamic leaf phenologyexplicit modeling of tree growth and dimensionstrunk diameter, height growth curve, rooting depth, bark thickness, crown base height, stem counts, basal areaaddition of a litter layer structureL, F and H litter layer depths and moisture transpiration from the H layer and mineral soilimplementation of fire modelingfire spread based on Rothermel’s mathematical modelfire mortality based on bark thickness 9

Implementation of RHESSys for Biscuit Brook10

Location of Biscuit Brook11

Overview of RHESSys worldfile creation12 Time series for minimum temperature, maximum temperature and daily precipitation are the minimum required climatic inputs. Soil and vegetation characteristics are specified via parameter definition (*.def) files. The GIS-based preprocessing step allows broad flexibility in partitioning the landscape (i.e., basins, hillslopes, micro-climatic zones, landscape patches).

Derivation of Biscuit Brook Hillslopes13

Detail of Biscuit Brook Hillslope Derivation14

Model Calibrationhydrological calibrationmodeled stream flow vs. actual stream flow datamodeled soil moisture vs. actual soil moisture datamodeled evapo-trans vs. actual evapo-trans datavegetation calibrationmodeled height or DBH growth vs. actual growthmodeled leaf area index vs. actual leaf area indexmodeled basal area per hectare vs. actual basal areamodeled biomass accumulation vs. actual accumulationsnow pack calibration (under development )15

(un-calibrated modeling results)Examples of RHESSys model output for Biscuit Brook16

Modeled streamflow 117

Modeled Streamflow 218

Modeled Streamflow 319

Vegetation Cover Comparison 120

Vegetation Cover Comparison 221

Terrain analysis example 122The western portion of the ridgeline and upslope region of Biscuit Brook with 2m contours. Steep outcrops surrounding relatively flatter terrain is prominent at this location, as indicated by the outlined areas.

Terrain analysis example 223Overlay of wetness index and the derived stream network (blue lines) suggests that low lying upslope areas (i.e., shelves) are often at the source of 1st order streams. Alternating soggy and dry conditions could lead to nutrient loading in 1st order streams, depending on vegetation type and status.

Snow Modeling24

Modeled Snow Pack SWE example 125 March 1999: spatial variability in mean monthly snow pack SWE is high but the mean monthly SWE quantity is low.snow water equivalent (SWE) mm

Modeled Snow Pack SWE example 226Jan 1999: spatial variability in mean monthly snow pack SWE is low but the mean monthly SWE quantity is high. snow water equivalent (SWE) mm

Summary/ConclusionsTake Home MessagesSimulations are sensitive to speciesSimulations are sensitive to precipitation patternSpatially adjusted snow modeling outputindividual components of snow pack loss Future WorkAdd additional local tree speciesModel calibrationExpand scale of modeling (e.g., Neversink basin)Use the calibrated model to investigate the effects of climate change on Catskill forests 27

AcknowledgementsThanks to my fellow CUNY postdocs for their encouragement and expertise.Thanks to DEP Modeling Group staff for their feedback and help finding necessary data.Special thanks to Larry Band for his continued support and guidance28