Advancing Water Resources Research and Management

An Online Near-Real-Time Crop and Weather Statistics for Texas Northern Plains

X. Chen, R. Srinivasan

Blackland Research Center

Texas A&M University System

Abstract

A near-real-time application of crop and weather statistics for Texas northern plains is developed online for crop yield and soil water simulation and prediction. It uses EPIC (Environment Policy Integrated Climate) model to simulate and predict crop and weather indicators: Potential Evapotranspiration (PET), Evapotranspiration (ET), Root Zone Soil Water content, Sediment, Biomass production, and Yield for nine locations in Texas northern plains. The weather data are downloaded daily from Texas northern plains PET Network, at Amarillo. Simulations are conducted daily based on newly downloaded weather data for three crops: corn, winter wheat, and grain sorghum. In addition to the time series tabular results, charting and mapping utilities are developed using Java and GIS (GRASS) technologies.

Key Words

Crop, weather, simulation, EPIC, GIS, Java, Texas, online, GRASS

Introduction

Rapid advances of computer technologies recently have revolutionized crop and environmental modeling. High-speed computers and widely accessible Internet make it possible for users to access remote-operated computations conveniently at home. For agriculture related decision-making, crop growth and its impact on environmental concerns such as soil and water quality are essential to our agricultural and environmental sustainability. Computer modeling and its predictive functions improve the understanding of our land prospectus. GIS (Geographic Information System) is a fast growing technology that lets you manage, maintain, manipulate, and analyze your spatial data efficiently. GIS mapping capability provides modelers with a visual presentation of spatial data and modeling results. Together with Internet, GIS empowers direct and easy simulation, computation and visualization on the Web. The objectives of this paper are to present an online near-real time simulation with crop and weather indicators: PET (Potential Evapotransportation), ET (Evapotransportation), Root Zone Soil Water, Sediment, Biomass, and Yield -- of Texas northern plains, and help users to determine the optimal crop production under common management practices and local soil and weather conditions. The application was developed using Perl, Java, HTML programming, and GRASS GIS techniques.

Site descriptions

Real time agricultural weather data from nine locations (Dalhart, Dimmitt, Etter, Farwell, JBF (Bushland), Morse, Perryton, Wellington, and White Deer), in nine corresponding counties (Dallam, Castro, Moore, Parmer, Potter, Hutchinson, Ochiltree, Collingsworth, and Carson) is recorded and managed through Texas A&M Agricultural Research & Extension Center at Amarillo, Texas (http://amarillo2.tamu.edu/nppet/netpet1.html) .

The area is totaled at about 68238.75 km². Agricultural practices include row crops, winter wheat, grasses, and cotton in some areas. Table 1 shows the information of the nine counties for annual average temperature, daily maximum and minimum temperature, annual average precipitation, and dominant soil types. All data except Dallam are from USDA National Weather & Climate Center averaged from 1961 to 1990. Data of Dallam County is from USDA Soil Conservation Service and Forest Service edited in 1975.

Table 1. Weather and soil information for the nine counties in Texas northern plains.

Model and data requirements

EPIC (Environment Policy Integrated Climate) is a comprehensive crop growth and environmental assessment model developed by USDA-ARS in the 1980's (Williams, et al., 1985). It simulates crop growth, soil, hydrology, water quantity and quality changes, nutrient cycling, erosion, sedimentation, pesticide fate, economic, and management practices such as tillage, fertilization, and pesticide application, drainage, irrigation, liming, pests, furrow diking, grazing, crop rotation, and their impacts on crop growth and environment sustainability. The EPIC model is capable of simulating most row crops, small grains, legumes, pastures, rice, casava, lentils, and pine trees, hundreds of soil types, and almost all processes involved in hydrology, nutrient cycling, erosion, management practices. EPIC is a continuous simulation model on a daily basis, which can be used for crop production and water resources management. It can also be used to assess and determine optimal management strategies for a wide variety of weather, crop, soil, and management practices. Effects of global weather/CO2 changes, nitrification, and volatilization submodels are among those newly added procedures in EPIC. The EPIC is capable of analyzing about 900 soils and 500,000 crop/tillage/conservation strategies across the U.S.

Data required for the EPIC model include parameters for crop, soil, hydrology, erosion, nutrient cycling, and operation schedule. Considering the purpose of this study and the complexity of the input data for the model, the only input required of users is the choice of crops. Three common crops are selected. They are corn, grain sorghum, and winter wheat. Cotton is a main crop in the southernmost counties, but due to the dominance of the other crops in this area, cotton is deleted in this application. Management strategies are defined in the operation schedule file in which most common practices are defined for each crop. The dominant soil type for cropland in each county location is used in the model. A homogenous soil within each location is assumed. Weather data used is either recorded field data sets or weather generated by computer from historical weather records.

GRASS GIS

GRASS (Geographical Resources Analysis Support System) (Shapiro et al., 1993) is a public domain raster-based GIS, vector GIS, image processing, graphics production, data management and spatial modeling system originally developed by the Environmental Division of the US Army Construction Engineering Research Laboratories in Champaign, Illinois, with enhancements and support now provided by the GRASS Research Group at Baylor University. Widely used by many military and nonmilitary agencies, GRASS GIS is used to collect, store, manipulate, analyze, and output geographic data.

Web interfaces

The web interface is developed using Perl, HTML, and Java programming and GRASS GIS technology accessed through http://srph.brc.tamus.edu/html/epic/html/epicappl.html. Figure 1 shows the schematic processes of the interface.

Weather data of both daily and hourly records is downloaded daily on the early morning and processed to extract necessary information for running the model. The downloaded data is that recorded two days ago. Preprocessor extracts daily weather parameters from the original daily and hourly data files. These parameters include maximum and minimum temperature, precipitation, solar radiation, relative humidity, and wind speed. Historical weather data from January 1, 1998 from the nine locations are also processed and used for simulation. Missing data sets will be developed using a built-in weather generator module in EPIC.

Figure1. Flowing of the program processes

To reduce online processing time, the EPIC model is run for all crop and location combinations once the preprocessing is done. Corn and grain sorghum are simulated as one-year continuous crops, while winter wheat is simulated across a two-year period. Post-processing involves extracting weather indicators: Potential Evapotranspiration (PET), Evapotranspiration (ET), Root Zone Soil Water content, Sediment, Biomass production, and Yield, from model output for both daily and monthly representations, constructing HTML files, and input files for Java applet charting and GIS mapping. A view of all output linkages is illustrated in Figure 2.

Figure 2. Page of links for output

Tabular data gives daily output of the model for the five weather indicators, as well as precipitation and irrigation for each crop on each location. They are distinguished through varied background colors. Weather data used to run the model in the current day is highlighted also. Outputs beyond the current day are predicted based on the simulated weather data.

Charts are created through Java applets modified from Jchart package from KL Group Inc. It is a client/server networking application utilizing Java 1.1 utilities and RMI (Remote Method Invocation) technology. They allow users to select chart types, title, legends, and other configuration parameters. Figure 3 shows one of the panels that allow choosing various feature values. The user can select either monthly or daily weather indicators to plot. For daily output, data is graphed for individual months, while monthly output is plotted for the whole year. Figure 4 is a chart of daily biomass production in April.

Figure 3. Panel for feature values selection.

Figure 4. Chart of daily biomass production.

Mapping soil moisture and biomass for each crop on nine locations is conducted through GRASS GIS technology. For either soil moisture or biomass mapping, parameter raster map layer, Texas county vector map layer, and site location layer are overlaid and labeled. It is then converted to GIF (Graphics Interchange Format) file format for display. Figure 5 is an example of .gif image generated by GRASS GIS. Legends on the bottom of the image indicate represented parameter values for specified colors. Unit for biomass is t/ha and cm for moisture, which is specified at the link to the image.

Figure 5. Map of soil moisture in Corn field.

For convenience, raw output results of the EPIC model are also listed on the bottom (see Figure 2).

Conclusions

This paper presents an online near-real-time application of crop and weather statistics for Texas northern plains. The application uses the EPIC crop growth model to simulate and predict crop and weather indicators of potential evapotranspiration, evapotranspiration, root zone soil water content, sediment, biomass production, and yield for nine locations in Texas northern plains. Weather data is downloaded daily from the Texas northern plains PET Network, at Amarillo. Simulations are conducted daily for three crops: corn, winter wheat, and grain sorghum, which are three main field crops in the Texas northern plains. The simulation provides potential users with plain text results, chart and maps developed using Java and GIS (GRASS) technologies. Easy-to-use interfaces and straightforward viewing of the results greatly facilitate user's understanding of the land performance and potentials for selected crops and locations. This application is a forerunner for potential extension to other areas to provide detailed crop and weather diagnostic estimations.

References

Williams, J.R., J.V. Putman, and P.T. Dyke, 1985. Assessing the effect of soil erosion on productivity with EPIC. Pp. 215-226. In Erosion and Soil productivity. Proc. Nat. Symp. Erosion and Soil Productivity. ASAE Publ. 8-85.

Shapiro, M., J.Westervelt, D. Gerdes, M. Larson, and K.R. Brownfield, 1993. GRASS 4.1 Programmer's Manual. U.S. Army Constructure Engineering Research Laboratory.

Symposium TOC	AWRA Home page