International Journal of Advancements in Research & Technology, Volume - PDF document

International Journal of Advancements in Research & Technology, Volume 3, Issue 1, January-2014 82 Copyright © 2014 SciResPub. IJOART International Journal of Advancements in Research & Technology, Volume 3, Issue 1, January-2014 81 Copyright © 2014 SciResPub. IJOART International Journal of Advancements in Research & Technology, Volume 3, Issue 1, January-2014 80 Copyright © 2014 SciResPub. IJOART International Journal of Advancements in Research & Technology, Volume 3, Issue 1, January-2014 79 Copyright © 2014 SciResPub. IJOART International Journal of Advancements in Research & Technology, Volume 3, Issue 1, January-2014 78 Copyright © 2014 SciResPub. IJOART International Journal of Advancements in Research & Technology, Volume 3, Issue 1, January-2014 77 Copyright © 2014 SciResPub. IJOART International Journal of Advancements in Research & Technology, Volume 3, Issue 1, January-2014 76 Copyright © 2014 SciResPub. IJOART International Journal of Advancements in Research & Technology, Volume 3, Issue 1, January-2014 75 Copyright © 2014 SciResPub. IJOART International Journal of Advancements in Research & Technology, Volume 3, Issue 1, January-2014 74 Copyright © 2014 SciResPub. IJOART Kharif 2009 Generation Standard Week 1 Accumulated GDD Egg Stage 2 Larval Stage 2 Pupal Stage 2 I (July 2) 27 59.59 1 22 27 28 190.56 - 13 18 29 316.25 - 5 10 30 446.24 - - 1 31 574.25 - - - II (Aug 7) 32 41.99 2 25 30 33 165.24 - 16 21 34 281.20 - 8 13 35 388.33 - 1 6 36 468.1 - - - Rabi 2009 I (Nov 7) 45 13.45 4 36 44 46 117.36 - 27 35 48 309.92 - 8 16 50 469.65 - - - II (Jan 2) 1 43.39 2 34 41 2 119.95 - 26 33 3 209.35 - 17 24 4 280.93 - 11 17 5 355.08 - 4 10 6 470.3 - - 1 1 Standard Week refers to one week in a calendar year, beginning as 1 in Jan and ending as 52 in Dec 2 Refers to the number of days remai ning for completion of the stage IJOART [14]Weiser,M., Brown, J.S1998. The Coming Age of Calm Technology,in: Denning, P.J., Metcalfe, R.M., Beyond Calculation The Next Fifty Years of Computing, Springer, New York, pp. 75[15]Wu, L., Damicone, J.P, Jackson, K.E., 1996. Comparison of Weather based Advisory Programs for Managing Early Leaf spot on Runner and Spanish PeanutCultivars. PlantDis., 80, 640645.[16]Baggio,A., 2005, Wireless Sensor Networks in Precision Agriculture, Retrieved From http://www.tudelft.nl/live/pagina.jsp?id=b66fc20e24c4 1f8aef51b000&lang=en. Last accessed on 20 June, 2012.[17]Bheenaveni, R., 2007, Agriculture in India Issues and ChallengesRetrieved From http://www.articlesbase.com/selfpublishing articles/agricultureindiaissueschallenges 203476.html, Last accessed on 20 June, 2012. [18]Dallal, G.E., 2005, Randomized (Complete) Block Designs, Retrieved From http://www.jerrydallal.com/LHSP/ranblock.htmLast accessed on 20 June, 2012[19]ikisan, 2000, Groundnut Disease Management, Retrieved From http://www.ikisan.com/Crop%20Specific/Eng/links/ap_grou ndnutDisease%20Management.shtml, Last accessed on 20 June, 2012[20]Seybold, S.J., Donaldson, S., 2004, Pheromones in Insect Pest ManagementRetrieved From http://www.unce.unr.edu/publications/files/ag/other/fs9841. , Fact Sheet 98Last accessed on 20 June, 2012. [21]Sivakumar, M.V.K., 2008, Scientia Agricola, Retrieved From http://www.scielo.br/pdf/sa/v65nspe/a02v65nsp.pdf Last accessed on 20 June, 2012[22]http://www.memsic.co [23]http://www.tinyos.net Table 4. WSNAdvisories Issued For Leaf Spot During Kharif & Rabi 2009 Kharif 2009 Crop Age 51 55 58 61 63 68 71 74 77 86 89 92 95 99 102 Leaf Spot Grade 0.1 0 0 0.4 0.8 0.5 0.95 0.75 0.6 1.35 1.35 1.6 3.25 3.1 3 LWI Advisory Yes No Yes Yes Yes No No No No Yes NA 1 NA NA NA Yes T/RH Advisory No No Yes Yes No No No No No Yes NA NA NA NA No Similarity between Models No Yes Yes Yes No Yes Yes Yes Yes Yes - - - - No Spray Fungicide based on WSN Advisory and Disease Grade No - No No No - - - - Yes - - - - No Rabi 2009 Crop Age 52 58 62 66 72 77 81 83 87 93 100 107 114 118 124 Leaf Spot Grade 0 0 0 3.3 3.3 2.7 3 2.3 5 4.7 5 4.3 5 5 4.7 LWI Advisory Yes Yes Yes Yes Yes NA NA NA Yes ND 2 ND No Yes NA NA T/RH Advisory No No No No Yes NA NA NA No ND ND No No No No Similarity between Models No No No No Yes - - - No - - Yes No No No Spray Fungicide based on WSN Advisory and Disease Grade No No No No Yes - - - Yes - - - Yes - - 1 Not Applicable 2 No Data Table 5. CalculationsUsing The WSN Based Leaf Miner Advisory Model IJOART Yield (Kg/Ha/RBD) 1400 1750 1650 1318 Cost Benefit Ratio 3 1.1 2.1 8 3.35 - 1 FP Farmer's Practice; CP Complete Protection; WA WSN Advisory; NP No Protection 2 Cost per spray per ha = 620/ - 3 Calculated with reference to NP treatment. Ratio of additional yield (@ 25/ - per Kg) and Fungicide Cost Successfully forewarns the disease outbreak. Moreover, the cost benefit analysis reveals that reduction in the peticide and fungicide usage is possible without affecting the yield in a significant way. Though the WSN costs have not been inclued into the cost benefit analysis, efforts are being made to miimize the number of sensors and increase the area of coverage to make the system economically viable. WSN provides nmerous opportunities of research in microclimate analysis, which is absent in the current agricultural scenario. Greater proliferation of WSN based systems will enable optimal straegies to be developed for better crop management. With Dminishing Yields and Pest & Disease menaces being a worlwide concern, WSN provides a posible solution to smarter agriculture.UTURE ORKWSN based Pest and Disease Advisories will be correlated with Remote Sensing data to investigate the possibilities of increasing the coverage area of advisory. Collaborative rsearch activities have been proposed with the National Rote Sensing Cetre (NRSC) to develop tools and models in this area. This would strengthen the possibility of developing a cost effective system, which can be afforded by the farming community. In addition, the optimal number of WSN motes for one hectare ofdeploment will be worked out as part of our future objectives.ACKNOWLEDGMENTSthank the Department of Electronics & Information Tecnology (DeitY), Ministry of Communications & Information Tecnology, Government of India, for their continual supporttowards research in Ubiquitous Computing. A word of apprciation towards the Central Research Institute for Dryland Agriculture for their collaboration and domain expertise, without which, this research would not have been fruitful. Finally, we thank the Centre for Development of Advanced Computing encouraging and motivating us during our work.EFERENCES[1]Brunette, W., Lester, J., Rea, A., Borriello, G., 2005. Some Sensor Network Elements for Ubiquitous Computing. Fourth Int. SympInf. ProcessSens. Netw., 388 392. [2]Burrell, JBrooke, T., Beckwith, R., 2004. Vineyard Computing: Sensor Networks in Agricultural Production. Pervasive Comput., IEEE, 3, 3845. [3]Butler,D.R., Wadia, K.D.R., Reddy, R.K., Das, N.D., Johnson, B., Kumari, M., Krishna Murty, K., Sreenivas, B., Srivastava, N.N., 2000. A weatherbased scheme to advise on limited chemical control of groundnut Leaf Spot diseasesin India. Exp. Agric., 36, 469478.[4]Das, H.P., DoblasReyes, F.J., Garcia, A, Hansen, JMariani, L, Nain, A., Ramesh, K., Rathore, L.S., Venkataraman, R., 2010. Weather and Climate Forecasts for Agriculture, in: World Meteorological Organization, Guide to Agricultural Meteorological Practices, EPublishing, Switzerland, pp. 51 to 557. [5]Jensen, R.E., Boyle, L.W., 1966technique for forecasting leaf spot on peanuts. Plant Dis. Report., 50, 810814. [6]Kumar, P.V., Stigter, K., Brunini, O., 2010. Agrometeorology and Groundnut Production, in: World Meteorological Organization, Guide to Agricultural Meteorological Practices, EPublishing, Switzerland, pp. 9 to 1024. [7]Panchard, JRao, S., Sheshshayee, M.S., Papadimitratos, , Kumar, S, Hubaux, J.P., 2008. Wirelesssensor networking for rainfed farming decision support. 2ACM SIGCOMM workshop on Netw. Sys. Dev. Reg., 31[8]Parvin, D.W, Smith, T.H, Crosby, F.L., 1974. Development and evaluation of a Computerized Forecasting method for Cercospora LeafSpot of Peanuts. Phytopathol., 64, 385[9]Puccinelli, D., Haenggi, M.2005. Wireless Sensor Networks: Applicationsand Challenges of Ubiquitous Sensing. Circuits Syst. Mag., IEEE, 5, 1931. [10]Seongeun, Y., Jaeeon, K., Taehong, K., Sungjin, A., Jongwoo, S., Daeyoung, K., 2007. Automated Agriculture System based on WSN. Int. Symp. Consumer Electron., IEEE, 1 [11]Shanmuganthan, S., Ghobakhlou, A., Sallis, P., 2008. Sensors for modeling the effects of climate change on grapevine growth and quality. Proc. 12World Sci. Eng. Acad. Soc. Int. Conf. Circuits, 315320.[12]Shanower, T.G., Gutierrez, A.P., 1993. Effect of emperature on Development rates, Fecundity and Longevity of the Groundnut Leaf Miner, AproaeremaModicella, in India. Bull. Entomol. Res., 83, 413419.[13]Weiser, MBrown, J.S.,1996. Designing Calm Technology. PowerGrid J., 1, 117. IJOART Fig. 7: Disease Incidence Grade vs Leaf Wetness Index depicted in Table 4, field observations were done at different crop ages. During the Kharif seson, it was observed that Leaf Wetness Index (LWI) & TemperatureRelative Hmidity Index (T/RH) values exceeded the threshold at 58, 61 and 86 days of crop age which coincided with initial disease appearance and scouted disease incdence grade(as shown in 7). However, fungicide spray was not recommended since the disease incidence had not crossed the acceptable threshold. The unprecented drought conditions could be linked to the slow onset of the disease. A precationary spray was done at the crop age of 86 days following the WSN Advisory. A dormant window of 14 days was maintained following fungcide spray and therefore, the models did not issue advsories during this priod (represented as ‘NA’ in Table 4).the LWI model was deployed to sense the micrclimate, the T/RH model with an approximation suggested by agricultural scientists [15] was deployed to sense the macrclimate. Comparing the advisories issued by both models, we may coclude that there is close similarity in the timing of the advisories. We also observe that the LWI model issued three additional favorable advisories while the T/RH model rmained unfavorable. This dissimilarity may be attributed to the effect of the mcroclimate over the macroclimate.TheRabi crop was sown early November 2009. The crop was raised under irrigated conditions (using sprinklers) and the build up of disease was high during the season. The diease progress had higher slope value in unsprayed plots. While the T/RH index values exceeded threshold only once during the entire season, the LWI index values were higher than the threshold (i.e. 2.3) for much of the season (Table 4). Fungicides were applied based on the LWI advisory and the Leaf Spot grade. Betweenthe two models, there was minimal similarity during the Rabi season. This clearly brings out the variance between the microclimate and the macroclimate. This variance may be attributed to the fact that sprinklers irrigate the Rabi season. Due to the nature of the sprinklers, though the humiity build up occurs at the crop canopy, there is not much change at the stanard height (i.e. 1.5 meters, where T/RH is sensed). The experiment during the Rabi 2009 season indicates that the microclimate introduces greater dynamism in our uderstanding of pest and disease life cycle and is therefore an important consideration for pest and disease forewarning.Leaf Miner during Kharif & Rabi2009are calculated using the WSN weather data following the first adult catch in the pheromone traps erected in the field. The model provided information about the current stage in the life cycle of the pest. It also aided the prediction of IInd Generation of the pest, which is considered critical for the crop. The prediction of growth pattern was achieved by intgrating the real time WSN data with normalized historical data from the agrometeorological observatory. Table 5 lists the pest growth pattern based on the accumulated DDs.Cost Benefit AnalysisCostbenefit analysis was conducted for individual Rando ized Block Designs (RBD) based on the number of fungicide treaments applied and yield obtained in that RBD for the Rabi Season 2009.Resultsof the analysis have shown that the cost benefit ratio was higher in RBDs, which followed the WSN advisbased treatments. From the Table 6, it is evident that RBDs with Complete Protection scheduled eight fungicide sprays, while WSN RBDs scheduled only four, without affecting the yield significantly. In comparison to the Farmer’s Practice, the WSN RBD scheduled an additional spray but provided signiicantly higher yield.ONCLUSIONSFromthe discussions we may conclude that the microclimate plays an important role in the growth and oubreak of pests and diseases. Having implemented and compared the results of one macroclimate advisory (T/RH Model) and one microclimate advisory (LWI Model) for Leaf Spot disease, we may conclude that duing the Kharif season (rain fed), both models performed similarly and could be used interchangeably. But the experiment conducted during the Rabi Season (sprinkler irrigation) clearly reveals that the macroclimate model fails to predict the disease outbreak since thehumidity remains unafected by the sprinkler irrigation while the microclimate Table 6. Cost Benefit Analysis of WSN Advisory Rabi 2009 FP 1 CP 1 WA 1 NP 1 Number of Treatments 2 3 8 4 0 IJOART threshold temperature. Calculated DDs serve as the base to predict whether the pest is in egg or larval stage and thus asists in timing pesticide sprays.Table1. Threshold Temperatures and Degree Days for Each Stage n The Leaf Miner's Life Cycle Stage Temperature Threshold (°C) Degree Days Egg 12.4 60 Larva 11.3 327 Pupae 14.7 72 Adult 3 202 Thefirst adult catch in Pheromone traps [20] sets the biofix date for the model. The GDD is thereafter calculated indicaing the arrival of the next larval stage based on the current Temperature values and normalized historical data. The hitorical data copliments the real time data for calculation of GDD and aids prediction of the growth pattern for the pest thereby leading to acurate control action.BSERVATIONS ESULTSField ObservationsKharif and Rabi 2008Kharif and Rabi 2008, prior to the deployment of the Agri System, the crop was maintained untreaed throughout the season to record the natural pest and disease incidence pattern. The dates at which the damage (caused by Leaf Mier) reached a peak stage are represented in the Table 2. The table depicts that the seond generation of the Leaf Miner pest occurs during a critical stage in the crop life cycle. The peak population of the pest during this period is critical to the yield.Table 2. LeafMiner Activity Records For Kharif And Rabi 2008 Seasons Season Sown Date Groundnut Leaf Miner Pheromone Trap Catch Field Population Initia tion Date Peak Date Initiati on Date Peak Date Kharif 2 8/06 31/7 (2 1 ) 27/8 (10 1 ) 15/7 (4 2 ) 8/9 (27 2 ) Rabi 31/10 12/12 (2 1 ) 27/1 (16 1 ) 11/12 (1 2 ) 6/2 (17 2 ) 1 Number of Adult Moths / trap 2 Number of Larvae/ 10 Plants the Kharif season, observations indicated that Leaf Spot disease initiation was triggered by Temperatures 25 C during the day and around 20 C at night, coupled with � 90% during the early morning observation and around 70% during the noon observations. The disease progress rate (slope) was highest duing prolonged wet spells. In the same year during the Rabi season significant disease incidence was not oserved.Kharif & Rabi, 2009to unprecedented drought conditions that prevailed duing the 2009 Kharif season, sowing was delayed to early July. Weekly assessment of the Leaf Spot diseasewas done,on 3 randomly selected plants from the entire plot, a week folloing the apearance of the disease. All the groundnut plants surrounding the Leaf Wetness sensor were inspected for diease incidence. The disease infection was gradedon a scale of 5 as metioned in the Table 3.WSN Advisory ObservationsLeaf Spot during Kharif & Rabi 2009Leaf Wetness and T/RH indices were calculated each day for the weather data stored in the RAS database. Advisories were sued after 50days of crop sowing. A favorable advisory and an acceptable grade of disease incidence (as mentioned in Tble 3) were considered for spraying fungicide.Table 3. CriteriaFor Calculating The Leaf Spot Infection Grade Leaf Spot Infection Rating/Grade Numb er of Lesions observed on randomly chosen plants Area Covered with infection (%) 1 5 2 05 20 1 5 3 20 50 5 20 4 50 - 100 20 50 5 �100 � 50 depicted in Table 4, field observations were done at different crop ages. During the Kharif season, it was observed that Leaf Wetness Index (LWI) & TemperatureRelative Hmidity Index (T/RH) values exceeed the threshold at 58, 61 and 86 days of crop age which coincided with initial disease appearance and scouted disease incidence grade (as shown in 7). However, fungicide spray was not recommended since the disease incidence had not crossed the acceptable threshold. The unprecdented drought conditions could be linked to the slow onset of the disease. A precautionary spray was done at the crop age of 86 days following the WSN Advisory. A dormant window of 14 days was maintained following fungcide spray and therefore, the models did not issue advsories during this priod (represented as ‘NA’ in Table 4). IJOART Gateway Subsystem (GS)TheGS is a bridge btween FS and RAS. The GS comprises a WSN Gateway and a RAS Inteface Unit. The WSN Gateway runs TinyOS2.x with routing and the dissemination compnents. The RAS Interface Unit is a single board computer, which runs the Linux OS. It is prgrammed to convert raw sensor data to standard engineering units and store on a local database. This weather data is perodically uploaded to the RAS through an Internet modem. The GS is currently eclosed in a woodenbox similar to the FS motes in order to protect it from sun and rain and is powered by mains, backed up by an Uninterrupted Power Supply (UPS) sytem. At present, a microcontroller based solar powered low power gateway is being designed and developed ihouse to operate in a standalone mode, without mains power. The dsign challenge restricts the average power consumption to 1 Watt, cosidering continuous operation due to networking requirements. The gateway is based on the ARM Cortex Mprocessor, running the CooCox Operating system and intefaces with a WSN mote and GSM modem.emote Administration System Fig. 6: Remote Administration System TheRAS hosts a web server and a database. Weather data from FS is stored on the database and is provided as input to data analysis and decision support advisory models. Regitered users receive weather based decision support advisories for Pest and Disease Forewarning as Short Message Service (SMS) messages. Fig6 illustrates the modules developed on theRAS as part of the uAgri system.Decision Support Advisory ModelsTwoweather based decision support advisory models for groundnut Leaf Spot and one model for groundnut Leaf Miner are implemented under the supervision and guidance of dmain experts from CRIDA. The Leaf Wetness Index (LWI) model [3] and TemperatureRelative Humidity Index (T/RH) model [8] forewarn Leaf Spot disease development. The dvelopment life cycle of groundnut leaf miner is modeled using Growing Degree Days (GDD) [12]. The foowing factors are considered in the model development.(a)The crop age is 130 days from sowing to harvest(b)Disease incidence is low in the initial 50 days.(c)A window of 14 days is maintained between fungicide spraysLeaf Wetness Index AdvisoryInfectioof groundnut by pathogens, causing early and late Leaf Spot diseases, is strongly influenced by accumulated Leaf Wetness spells each day. The infection is severe and spray is advised when the cumulative 7 day Wetness Index (WI) eceeds a threshold of 2.3and the disease incidence exceeds 10%. Leaf Wetness Index (LWI) for a day is computed from Leaf Wetness hours. If Wetness Hours (WH) in a day is 20 or less, then WI is set to WH/(1)andwhen greater than 20, the WI is derived from the expre175*(2)T/RH Index AdvisoryThemodel renders a dayday forewarning for groundnut Leaf Spot assuming the availability of hourly observations of perature and Relative Humidity (RH) for previous five days. Number of hours with RH and the minimum Temperature during those hours are used for calculating T/RH index. Hours of RH are limited between 2 and 20, while the Minimum Temperature is limited between 62and 80T/RH index is derived from the graph plotted bween number of Hours with RH and minimum Teperature [5]. The model is implemented with an approximtion suggested by similar experiments conducted by researcers at Oklahoma State University [15]. The RH threshold is therefore reduced to 80% (measured at 1.5 meters) rather than 95% (at canopy) as suggested in the original citation. The T/RH model therefore contributes to the macroclimate analsis of the Leaf Spot disease.Growing Degree Day ModelDegree Day (DD) corresponds to the difference of one dgree between mean temperatures each day, and a reference Temperature. The reference Temperature is a threshold Teperature that governs the development of the pest. This value varies with different stages of the pest life cycle. Each stage requires the accumulation of a fixed number of DDs for transtion to the next stage of development. The date to begin accmulating DDs, known as the biofix, varies with the species. A leaf miner copletes its life cycle in 660 DDs (Table 1), above IJOART multihop network to route sensed information to the GS. A single mote is deployed to sense the macroclimate and is intefaced with coarsely varying sensors like Solar Radiation, Raifall and Wind Speed & Direction.FigDeployment at Hyderabad, APFig. 3: Deployment at Ananthapur, AP TheWSN deployment is carried out in open farm condtions, in 3 different locations namely Hyderabad (Fig2), Aanthapur (Fig. 3) and Kadiri (Fig4). In the Hyderabad dployment, 6 micro climate motes and 1 macro climate mote are ployed in an area of 4 acres, while in Ananthapur, which is considered as the groundnut belt of India, 25 micro climate motes, developed inhouse, are deployed in an area of 10 hetares. TheKadiri deployment has 10 motes, covering 20 acres of land. Presently,the motes are encased in wooden boes with air vents popularly known as Stevenson's screens, with the Temperature & Relative Humidity sesors located within. The mote unit is powered by a battery that is charged by a solar panel, making the setup ttally standalone. FigDeployment at Kadiri, APFig. 5: Randomized Block Designs Thefield is divided into blocks (Fig5), and each block is administered pest and disease treatments based on six stratgies (T1 T6). This representation, referred to as Radomized Block Design (RBD) [18], aids scientific treament analysis of the crop, leading to cost benefit analysis. IJOART the Decision Support Advisory Moels developed as part of the research. The Observations and Results are analyzed in Section 4, while the Conclusions are summarized in Section 5, lowed by the Future Work, listed in Section 6ELATED ORKAn experiment in field crop production, Lofar Agro, deals with fighting phytophthora in a potato field. To monitor Reltive Humidity, Temperature and Leaf Wetness, which are iportant indicators to the development of the disease, the potto field was instrumented with wireless sensors [16]. In anoter experiment Atomated Agriculture System (A2S), a WSN was deployed in greenhouses with melon and cabbage in Dongbu Handong Seed Research Center. A2S was used to monitor the growing process, and control the illumination within the greenhouses [1]. A prototypeconsisting of a wirless network of groundsensors periodically recordsoil moiture, temperature, humidity and atmosphic pressure in the field environment. Data sensed was used for forecasting, forewarning and ultimately to increase productivity [7]. Simlarly, a WSN was deployed to monitor weather and enviromental conditions that affect the phenological stages of varous grapevine varieties in different countries [11]. Intel rsearched WSN in vineyards and worked out methods of better managment based on the ethnographic research [2].YSTEM RCHITECTUREpart of the National Initiative in Ubiquitous Computing, a pilot, WSN based, Ubiquitous Agriculture (uAgri) system is developed and deployed in groundnut research farms at Hderabad, India. The project is a collaborative research effort between the Centre for Development of Advanced Computing DAC) and the Central Research Institute for Dryland Agrculture (CRIDA). The aim of the project is to investigate the effect of microclimate on pests and diseases in groundnut and provide forewarning advisories. DescriptionThesystem architecture (Fig1) comprises three components namely Farm Site (FS), Gateway Subsystem (GS) and Remote Administration System (RAS). The FS consists of WSN motes ployed in the groundnut field. The GS aggregates sensed weather data from the FS and stores it on the RAS. The RAS utilizesWSN weather data for data analysis and decision suport advisory models. The detailed description of the system architeture is as follows. Farm Site (FS)TheWSN motes in the FS include IRIS motes, purchased from Memsic Inc. [22] as well as motesdeveloped inhouse, based on CC2430 SoC and MSP430. The motes are prgrammed to run a TinyOS 2.x [23] application for sensing and multihop routing of sensed data to the GS. The motes also provide a feature of control information dissemination, that enables an user to cofigure parameters like sensing intervals, query the mote for its health specific information and, also request shot sensor iformation. In normal mode of operation, the parameters are sensed once in every hour according to the domain requirements, but can also be confiured to sense at intervals as low as 10 seconds. FigAgri System ArchitectureWSN data packets are created whenever a sensor is sapled, based on its configured periodicity. These packets are multihopped to the GS using the Collection Tree Protocol (CTP), developed in TinyOS. A number of other routing prtocols like TinyAODV, Multihop LEPS and Static Routing were also integrated and tested during the field trials, but CTP was chosen for its satisfactory performance. In order to achieve ultralow power network operation, which is a wellknown challenge in outdoor deployments, it is mandatory that the mote radio be dutycycled in a coordinaed manner. Timesynchronization algorithms are therefore required, to perform coordinated network sleeping. The costrained resources on a mote due to its low processing power and memory, curtail this process. Further, integration of these algorithms with existing routing algorithms proves challening. A Time Division MultipleAccess (TDMA) based aproach for network coordination has been developed inhouse, which prvides both multihop routing and control information dissemination. Simulation results have proven that ultralow power consumption is achieved through 1 pecent radio duty cycling. This algorithm is currently being intgrated into the uAgri system for field trals.Two classes of motes are deployed, which enable micrclimate and macroclimate monitoring. The microclimate sening motes are deployed in the crop canopy area and are intefaced with sensors like Temperature, Relative Humidity,Leaf Wetness and Soil Moisture & Temperature. Tempeature and Relative Humidity is measured at a domain specfied standard height of 1.5 meters and Leaf Wetness, Soil Moisture & Temperature is measured at crop canopy. These motes are searated by a distance of 100 meters and form a IJOART Wireless Sensor Network based Forewarning Models for Pests and Diseases in Agriculture A Case Study on GroundnutSantosh Sam Koshy, Yesho Nagaraju, Sowjanya Palli, Y. G. Prasad, Naveen PolaTeam Embedded, Centre for Development of Advanced Computing, Hyderabad, IndiaDepartment of Entomology, Central Research Institute for Dryland Agriculture, Hyderabad, IndiaEmail: santoshk@cdac.inynagaraju@cdac.insowjanyap@cdac.inygprasad@crida.ernet.in, naveen_pola@yahoo.co.inABSTRACT In Agriculture, microclimate plays an important role in the growth and outbreak of pests and diseases. Wireless Sensor Networks (WSN) enables the acquisition of both microclimate and macroclimate weather data from agricultural farms thereby facilitating new insights into the cropweather dynamics in and around the crop canopy. In this paper, we present the results of an open farm deployment of WSN for the groundnut crop, with emphasis on weather based Pest and Disease Management. Having implemented two decision support advisory models for a groundnut disease, and one model for a groundnut pest, we highlight the importance of the microclimate over the macroclimate. We also discuss the cost benefit of the WSN based Advisory over the farmer's practice and other standard practices. KeywordsUbiquitous Computing, Wireless Sensor Networks, Micro Climate, Leaf Spot, Leaf Miner, Groundnut, Randomized Block DesignsNTRODUCTION Ubiquitous Computing (UbiComp), the third wave of compting, follows the eras of mainframe and personal computing [13]. Ubicomp is considered as the age of Calm Technology [14] where technology recedes into the background while redering its supportive services in an unobtrusive way. Wireless Sensor Network (WSN) is one step in this direction, enabling Ubiquitous Computing to proliferate our daily spaces [1]. Conceptually speaking, WSN combines various technologies like Sensing, Prcessing and Wireless Communication into a system architecture that facilitates the interfacing of the physcal world with Cybespace [9]. Agricultural practices need to address problems like clmate change, land infertility, diminishing yields, and rampant pest outbreaks [17]. The knowledge of weather helpsin adressing a few of these problem areas satisfactorily. Automatic Weather Stations (AWS) measuring parameters like Temperture, Relative Humidity, Rainfall, Solar Radiation and Wind Speed & Direction, provide macroclimate information [4]. Aricultural research also emphasizes the need for understaning microclimate within the crop canopy by measuring prameters like Leaf Wetness, Soil Moisture & Temperature and Canopy Temperature & Humidity [21]. Groundnut in India, is cultivated in low to moderate infall zones [6]. The crop age typically varies between 90130 days. It is grown in two seasons Kharif and Rabi during a caendar year. The Kharif season (JuneSeptember) is characteized by rain fed agriculture and during the Rabi season (NvemberFebruary), fields are irrigated. Investigations reveal that Leaf Miner is considered to be one of the major pests, of the groundnut crop. Temperature plays an important role in the pest’s growth, as the pest requires the accumulation of fixed amounts of heat units, to pass from one stage to the next of its life cycle [12]. The condtions favorable for leaf miner growth are, long dry spells resulting in high Temperature and low humidity [6].Groundnut is also prone to attack by numerous diseases. Among fungal fol iar diseases, only a few are economically portant in India such as Leaf Spot (early and late) and Rust. These are widely distributed and can cause yield losses in suceptible genotypes to the extent of 70% [19]. Weather condtions congenial for occurrence of early and late Leaf Spot are rainfall, moisture causing leaf wetness and temperature [6]. WSN facilitates the aggregation of microclimate infomation from agricultural fields by installing sensors within the crop canopy at various locations in the field. This microclimate information supports the analysis of various factors that inflence crop and pest growth thereby aiding the development of decision support advisory models. The decision support advsories help farmers to make better decisions in crop managment. WSN’s sphere of influence in agriculture, encompass areas like irrigation, pest & disease management, drought analysis & early warning, precision farming etc. The paper is divided into six sections. Section 1 briefly rviews some agricultural experiments using WSN. Section 3 scribes the Ubiquitous Agriculture System Architecture and IJOART