Location: Agricultural Systems Research2007 Annual Report
1a. Objectives (from AD-416)
Objective 1.Develop, implement, enhance, and maintain an object modeling system (OMS) and a library of modules for building agricultural system models at field to watershed scales for a variety of applications. [Contributes to Problem Area #1, Effectiveness of Conservation Practices, Product #5 of the new National Program (NP) 201 Action Plan (FY 2006 - 2010)] Objective 2.Develop, verify, and evaluate field to watershed modeling tools and techniques that quantify environmental outcomes of conservation practices in major agricultural regions, including modeling and decision aids for drainage water management systems. [Contributes to Problem Area #1, Effectiveness of Conservation Practices, Product #5 and Problem Area #3, Drainage Water Management Systems, Product #4 of the new NP 201 Action Plan (FY 2006 - 2010)] Objective 3.Develop improved space-time scaling and model parameterization approaches for landscape processes in new agricultural system models from field to watershed scales. [Contributes to Problem Area #1, Effectiveness of Conservation Practices, Product #5 of the new NP 201 Action Plan (FY 2006 - 2010), and to Goal 1.7.2 of NP 201 to develop methods to determine input model parameters, values, and state variables for multiple scales to account for the effect of management practices].
1b. Approach (from AD-416)
Objective 1. Hypothesis: The OMS framework can be used to develop customized, modular field to watershed ag system models with interchangeable components for assessing the effects of conservation practices. Experimental Design: OMS represents an ARS-led effort in partnership with the NRCS, USGS, and university collaborators (e.g. CO State University). Enhancing OMS functionality includes the development of improved capabilities for: 1) model building 2) code testing, data connectivity, and database integration; 3) geospatial output visualization and model parallelization; and 4) uncertainty, sensitivity analysis and parameter estimation. Objective 2 Hypothesis 2-1: A new prototype regionalized model can provide improved estimates of the effects of conservation practices on environmental responses at the field to watershed scales. Experimental Design: The overall goal is to develop an OMS-based modular simulation model with interchangeable components that can address regional soil and water conservation and water quality need from field to watershed scales. Specific task areas for Objective 2 are: 1) Identify regions and define process modules for a selected regional area; 2) Obtain needed scientific model components; 3) Develop a prototype regionalized watershed model and perform a preliminary evaluation; 4) Modify existing modules or identify and develop additional modules; 5) Evaluate the prototype watershed model with various conservation practices; and 6) Transfer the prototype model to NRCS. Hypothesis 2-2: An agricultural systems model, RZWQM2, can simulate and quantify the effects of BMPs under tile drainage for different Midwest climate and soil conditions. Experimental Design: In a collaborative research effort with the National Soil Tilth Laboratory (Ames, IA). Field experiments will be conducted in Iowa. RZWQM2 will be used to quantify controlled drainage and cover crop effects on drainage volumes, nitrate losses in drainage flow, and crop growth. Objective 3 Hypothesis: Soil, water and plant properties can be scaled over space and time to identify scale-appropriate behaviors and model parameters across agricultural landscapes. The resulting perameters can be used to improve the modeling of spatial interactions between land areas containing differential management and conservation practices. Experimental Design: The prototype regionalized watershed model will be used to assess the propogation of uncertainty in model structure, parameter values, and inputs to water quantity and quality effects up to watershed scales. Scale-dependence and uncertainty of model parameters will be evaluated as follows: 1) Characterize the spatial and temporal variability of measured system variables in the prototype watershed model; 2) Relate key model parameters to spatial surrogates; 3) Generate high resolution inputs to detailed process modules and upscale the results; determine effective parameter values over the range of scales of interest; and 4) Quantify parameter uncertainty and its impacts on model output uncertainty using a suite of object-based tools developed for parameter estimation.
3. Progress Report
OMS was enhanced to better support the visual assembly of spatially distributed models for different space-time scales. Database and GIS data auxiliary components were developed that allow connection to NRCS and CEAP data warehouses for access to spatial and non-spatial climate, soils, water quality and quantity, and management data sets. A version of the Shuffled Complex Evolution (SCE) parameter estimation algorithm was integrated into OMS. OMS was enhanced to connect to the USDA Colab (Collaborative Software Development Laboratory) web-based project management environment for module library hosting, code tracking and version control. Twenty-six preliminary modeling regions in the U.S. based on the NRCS concept of Land Resource Regions (LRRs) were identified. For developing a prototype regionalized watershed model in OMS, the Midwest LRR region was selected. Some of the needed scientific components for the regionalized prototype watershed model in key process areas such as water balance, nutrient cycling, soil erosion, and plant growth and development were obtained from legacy models such as RZWQM, WEPP, PRMS, and the European watershed model J2000. We are also currently incorporating components of the SWAT model into OMS as part of an initial plan to update the SWAT model code base to a more modular, component-based structure. This will help facilitate integration of SWAT components into OMS and set the stage for further enhancement and science improvements. In addition, new structural linkages to components of the CONCEPTS and REMM models were investigated to enhance the ability of the prototype regionalized watershed model to improve simulations of the dynamics of water and sediment transport in channels and riparian areas, respectively. New empirical scaling relationships were investigated based on continuing experimental on-farm spatial and temporal data being collected in Colorado and elsewhere to assess variability over nested scales and across multiple landscape positions. Spatial data collection near Ault, Colorado for a wheat-fallow system includes a meteorological station plus precipitation; temperature above canopy and in soil; spatial crop yield; plant emergence, development, biomass, and leaf area; soil bulk electrical conductivity; soil texture; soil; infiltration; runoff at edge of field; and remotely-sensed images. Data accuracy and quality also were evaluated, e.g., the need to improve estimated soil water content using capacitive sensors led to advances in sensor characterization. Additional work was performed in the context of an ARS initiative on sensor technology to address temperature sensitivity of dielectric sensors. Recent work on estimating soil properties that control micro-environments is being used to relate soil hydraulic parameters of different soil textural classes to their pore-size distribution index. Thus, we are continuing to investigate how basic soil physical parameters that are measured spatially can be used to estimate soil hydraulic properties needed in a physically based model. Similarly, we are exploring other surrogate variables for estimating and scaling key soil and plant parameters.
Development and Enhancement of the Object Modeling System (OMS). The OMS application programming interface (API) was enhanced to allow the representation of different space-time domains and data structures (e.g., trees and grids). As proof of concept, the European J2000 object-oriented watershed model was integrated into OMS using the new API. OMS was also extended to transfer and retrieve science modules to and from the OMS Component Library residing under the USDA Colab (Collaborative Software Development Laboratory) project management environment. J2000 modules were then used to evaluate the new Component Library linkage. In collaboration with USGS, the LUCA Java-based software package for model calibration was integrated into OMS. LUCA implements the Shuffled Complex Evolution (SCE) algorithm for model-independent parameter calibration. In addition, a prototype OMS component that allows the visualization and manipulation of geospatial data is currently being developed using NASA World Wind geospatial technology, and a journal paper describing space-time data structures in OMS is currently under preparation. [Contributes to problem Area #1, Effectiveness of Conservation Practices, Product #5 of the new National Program (NP) 201 Action Plan (FY 2006-2010)] Development of Crop Growth Modeling Tools. PhenologyMMS Version 1.2 was released. This computer program can be used as a stand-alone tool or incorporated into existing crop simulation models and decision support tools to simulate changes in multi-crop phenology as a result of varying levels of soil water availability. Version 1.2 provides the complete developmental sequences of the shoot apex correlated with different developmental events for winter and spring wheat, winter and spring barley, maize, proso millet, hay millet, and sorghum. The program is available on CD or can be downloaded via the Internet. A stand-alone plant growth model derived from the WEPS plant growth model has been developed and initially tested for corn and wheat across a range of environments. The Unified Plant Growth Model (UPGM) merges different versions of the EPIC-based plant growth model that exist in many agricultural system simulation models and decision support technologies (e.g., GPFARM, WEPP, WEPS, SWAT, and ALMANAC). Using this stand-alone foundation, improvements in simulating plant growth have been made. For example, a beta version of new phenology and seedling emergence simulation code from the PhenologyMMS project was initially tested in UPGM and GPFARM. Further testing will be necessary before the code is ready for final incorporation. Extensive testing of the GPFARM plant growth model was conducted in the past year, and thorough evaluation of the harvest index approach for calculating yield is underway. This accomplishment represents milestones from the previous CRIS project (Scaling and Modeling Space-Time Variability of Landscape Processes to Enhance Management) continued under this new CRIS project. [Contributes to Problem Area #1, Effectiveness of Conservation Practices, Product #5 and Problem Area #3, Drainage Water Management Systems, Product #4 of the new NP 201 Action Plan (FY2006-2010)] Spatial Data for Model Scaling and Parameterization. A detailed data collection field experiment was established at the Drake Farm, Ault, CO. Data collected include high-resolution (5m) elevation data; soil samples for bulk density, texture, and gravimetric water content, soil water content and temperature at various depths across strategic landscape positions, soil nutrients; crop yield (from a combine yield monitor); surface water runoff; and plant measurements such as emergence, LAI, and phenology. Soil water infiltration measurements were collected at 150 points in clustered patterns at different landscape positions. Soil samples were collected 2 days after steady infiltration measurements at two depths. These data will be used to estimate soil water retention and hydraulic properties using “functional normalization” in comparison with infiltration rates. On-farm field measurements of landscape variables and processes have been analyzed for their spatial and temporal behaviors, and methods for spatial analyses have been developed and tested to help classify land areas and guide spatially distributed sampling and simulation efforts. This research has continued into the new project to test scaling and parameter estimation methods, effective parameter estimation concepts, and parameter sensitivity in distributed models. Rigorous scaling will necessarily be limited to “sensitive” processes and parameters given limited resources, but we aim to identify dominant processes at different scales and scale the relevant parameters. This accomplishment represents milestones from the previous CRIS project (Scaling and Modeling Space-Time Variability of Landscape Processes to Enhance Management) continued under this new CRIS project. [Contributes to Problem Area # 2.4: Site Specific Technologies to Conserve Water, Nutrients, and Energy, Product #7 of the new NP 201 Action Plan (FY2006-2010); Problem Area #6 Water Quality Protection Systems, Product #3; and Goal 1.7.2 of NP 201 to develop methods to determine input model parameters, values, and state variables for multiple scales to account for the effect of management practices.] RZWQM Model Evaluation of Cropping System Management Practices. Including winter cover crops such as winter rye in corn-soybean rotation is one of the more promising practices to reduce nitrate loss from tile drainage system without negatively affecting production. A calibrated RZWQM-DSSAT hybrid model was tested for simulating the effects of cover crop versus no cover crops on nitrate leaching losses in subsurface drainage water under a corn-soybean rotation. Field experimental data collected over several years in Boone County, IA, with an application rate of 225 kg N ha-1 in corn years, was used for model evaluation. Average observed and RZWQM simulated flow weighted annual nitrate concentration (FWANC) in subsurface drainage water for the cover crop treatments from 2002 to 2005 were 8.7 and 8.6 mg L-1, compared to 22.1 and 17.2 mg L-1 for no cover crop (resulting in observed and predicted reductions of 61% and 50%, respectively). Simulations based on various N fertilizer application rates showed that annual FWANC in subsurface drainage water dramatically increased with the increases in the N application rates. With a cover crop, average FWANC increased from 3.1 mg L-1 at the 0 kg N ha-1 rate to 13.1 mg L-1 at 250 kg N ha-1 rate. In comparison, a corresponding increase from 6.1 mg L-1 to 21.1 mg L-1 was observed without a cover crop present. [Contributes to Problem Area #1, Effectiveness of Conservation Practices, Product #5 and Problem Area #3, Drainage Water Management Systems, Product #4 of the new NP 201 Action Plan (FY2006-2010)] Watershed Data System Released to CEAP Research Team. Comprehensive, long-term data from diverse watersheds are needed for hydrologic and ecosystem analysis and model development, calibration and validation. To support the Conservation Effects Assessment Project (CEAP) in assessing environmental impacts of USDA conservation programs and practices, researchers and staff from multiple ARS locations (El Reno, OK; Columbia, MO; Beltsville, MD; Ames, IA; Fort Collins, CO) developed a web-based data system: Sustaining the Earth’s Watersheds, Agricultural Research Data System (STEWARDS). The data system organizes and documents soil, water, climate, land-management, and socio-economic data from multiple agricultural watersheds across the US and allows users to search, download, visualize, and explore data. Now being beta-tested by the CEAP research team, when released to the public STEWARDS will facilitate: 1) researchers in obtaining ARS’ long-term data for hydrological studies; 2) modelers in retrieving measured data for model calibration and validation; and 3) watershed managers and a wide array of partners and stakeholders in accessing long term data to support conservation planning and assessment. Anticipated benefits include preservation of data, increased data use, and facilitation of hydrological research within and across watersheds with diverse collaborators. [Supports NP 201, Component 1: Agricultural Watershed Management].
5. Significant Activities that Support Special Target Populations
On-farm spatial data collection and transfer of information to “small farms” are enabled under the MOU with the Drake Farm near Ault, Colorado, which is a family owned and operated business. Information from this cooperative work has also been shared with other owner/operators of small farms through meetings of the Young Farmers Association.