1a. Objectives (from AD-416)
The objectives are to: (a) evaluate the utility of climate variations and forecasts for agriculture and resources conservation applications; (b) develop risk-based decision tools that take into consideration climate variations and forecasts for practical decision applications in agriculture and natural resource management; and (c) demonstrate climate-related decision and application opportunities for a livestock grazing enterprise and a reservoir water-level managment plan. The guiding principle underlying this project is the bridging of the gap between emerging climate knowledge and application of climate information to problem-solving by developing decision tools for real-life applications that meet the requirements of producers and resource managers.
1b. Approach (from AD-416)
Decadl-scale climate variations are identified by a trend analysis of historical climate data published by NOAA's national Climate Data Center, and the utility of seasonal climate forecasts by NOAA's Climate Prediction Center are evaluated in terms of forecasted and observed departure from average conditions. Statistical characteristics of climate variations and forecasts are quantified in terms of basic districution statistics and probability of exceedance cures (POE). Associated weather outcomes are developed using a weather generator, which will drive selected crop and hydrologic models that simulate climatic impacts on forage production and natural resources. Collaborating producers will capture their decision process using a journaling approach which will identify critical decision variables for which POE curves will be developed. The POE curves will reflect the risk and uncertainty of forecasted impacts and represent the basic decision information for decision makers. Two case studies, the management of a fall forage-grazing system in central Oklahoma and water-level regulation for Lake Texoma reservoir, will be used to demonstrate the management protential offered by climate variations and forecasts.
3. Progress Report
Objective 1. The gridded weather data produced by the PRISM Climate Group and advertised as the "USDA's official climate data", include estimates of monthly precipitation every 4 km across the contiguous United States. These climate data appear to be useful for the development of climate-informed decision support for agriculture. However, the accuracy of the PRISM precipitation estimates had not been evaluated at very small spatial scales using independent data. Historic daily precipitation data collected at 8 closely located rain gauges in central Oklahoma was digitized, quality controlled, and summed into monthly values. These data were then compared individually and collectively to the collocated PRISM estimates. Significant differences were revealed, suggesting that the PRISM data is not accurately representing many of the larger precipitation events. Analysis of possible reasons for the differences continues. Objective 2. Oklahoma agricultural production, income statistics, and historical climate records were reviewed to identify scenarios of climate, market and initial soil water conditions that favored either grain production, beef production, or grain and beef production. Computer simulations of winter wheat grazing and beef and grain production were conducted for dry, average, and wet climatic conditions. The results of these simulations and the agricultural income statistics were interpreted to provide management guidelines that optimized income under varying climatic and market conditions. Objective 3. Generation of long synthetic daily weather records for computer simulation of crop yield, soil erosion, sediment and nutrient movement, and downstream water quality and sedimentation is common practice. ARS has developed several synthetic weather generators for agricultural applications. The random aspect of the weather is modeled by use of random numbers that are assumed to be uniformly distributed. However, a set of around 100 computer-generated random numbers, as commonly used for weather simulation, do not always adhere to the uniform distribution expected by the weather model. Under unfavorable conditions this can lead to an undesired bias in the synthetic weather. A novel technique has been developed to constrain sets of random numbers to closely adhere to the uniform distribution, thereby ensuring consistency between expected and actual distribution of random numbers. Such consistency between expected and actual distribution of random numbers increases the quality of the synthetic weather data and improves our ability to reproduce characteristics of the historical weather being simulated.
1. Fertilizer application rate optimized for forage production under various climate and market scenarios. In Oklahoma, it is common practice to graze winter wheat in the fall and winter for beef production, and to harvest the wheat for grain at the beginning of summer. However, managing this forage-crop system is quite challenging, as the tradeoff relationships between beef and wheat production are complex. Researchers at the ARS Grazinglands Research Laboratory in El Reno, Oklahoma, showed that earlier initiation of grazing is generally more profitable for the forage-crop system, irrespective of climate, because a longer grazing season favors profitability of the beef production. General guidance for optimal fertilizer rates along with their likelihoods of profitability was derived based on precipitation amounts between Aug. 1 and Feb. 28 during 1909-2007. Optimal fertilizer rates for dry summer and winter should be much less than the current average application rates, and doubled for a wet summer and winter. The findings help farmers make informed decision on how much fertilizer to apply at the end of winter based on preceding precipitation total.
5. Significant Activities that Support Special Target Populations
The laboratory partnered with ARS laboratories at Lane, OK; Booneville, AR; and Fayetteville, AR, to develop and staff an exhibit for the Know Your Farmer, Know Your Food Conference and Gala Dinner in El Reno, Oklahoma, in July 2010. The Conference was attended by over 200 persons, the majority being small- to mid-sized farmers, institutional food professionals, and extension or outreach specialists. The project's capacity in climate variability and forecast applications for decision support was included in the exhibit.
Garbrecht, J.D., Schneider, J.M., Brown, G.O. 2007. Soil water signature of the 2005-2006 drought under tallgrass prairie at Fort Reno, Oklahoma. Proceedings of the Oklahoma Academy of Science. 87:37-44.