2012 Annual Report
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.
This research project began 24 Nov 2009 and terminated 10 Mar 2012. Progress for the period 1 Oct 2011 through 10 Mar 2012 is reported under the new project 6218-11130-005-00D "Adapting Soil and Water Conservation to Meet the Challenges of a Changing Climate".
Life-of-the=project progress for project 6218-11130-004-00D for 24 Nov 2009 through 30 Sept 2011 is below.
The PRISM Climate Group disseminates estimates of monthly precipitation every 4 km across the contiguous United States. The accuracy of the PRISM precipitation estimates had not been evaluated at very small spatial scales. Such validation is necessary for field-scale agricultural applications of the precipitation estimates. Precipitation estimates were compared to actual daily precipitation measurements taken at 8 rain gauges in central Oklahoma. Significant differences were revealed, suggesting that the PRISM data is not accurately representing many of the larger precipitation events.
Agricultural production and income statistics and historical climate records were evaluated to identify various combinations 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 identified management guidelines that optimized income under varying climatic and market conditions. Also, the decision support potential of climate statistics and initial soil moisture at winter wheat seeding time was explored for a winter wheat-cattle grazing enterprise in Oklahoma. An enterprise budget was developed that included estimation of profit/losses and risk due to inclement weather and/or low initial soil moisture at seeding time.
The utility of the NOAA's seasonal climate forecasts for decision making was found to be of limited value for managing wheat-cattle production systems in Oklahoma. An alternative forecast method was explored. The method is based on the assumption that similarities can be found between the statistics of a forecast year and a year in the historical record. The observed weather of a year with statistics similar to the forecast is taken as the forecast year. Preliminary results showed that there were no recognizable patterns at the window time-scales of 30, 60, 90, and 120 days for both daily and monthly precipitation.
The unpredictable character of the weather is simulated in computer models by use of random numbers that are assumed to be uniformly distributed. Tests have shown that under unfavorable conditions a departure from the uniform distribution 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. 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.
Increasing profits by managing fertilizer application in terms of climate conditions. Researchers at the ARS Grazinglands Research Laboratory in El Reno, Oklahoma, showed that earlier initiation of grazing is generally more profitable for dual use wheat, irrespective of climate, because a longer grazing period favor profitability as the fixed total costs per steer are spread over more weight gain. General guidance for optimal nitrogen rates along with their likelihoods of success was derived based on precipitation amounts between Aug. 1 and Feb. 28 during 1909-2007. Optimal nitrogen 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 nitrogen to top dress based on preceding precipitation total.
Synthetic weather generators benefit from new random number generation method. The unpredictable character of the weather is modeled by use of random numbers. Tests were conducted to determine if computer-generated random numbers adhere to distribution properties set forth by the weather model. It was found that under unfavorable conditions a departure of the random numbers from the uniform distribution can lead to an undesired bias in the synthetic weather. A novel technique was developed to ensure random numbers closely adhere to the uniform distribution thereby ensuring consistency between expected and actual distribution of random numbers. The new random number generation increased the quality of the synthetic weather data and improved our ability to reproduce characteristics of the historical weather being simulated.
PRISM precipitation estimates do not live up to expectations. ARS researchers in El Reno, Oklahoma, compared the gridded monthly precipitation estimated using the PRISM climate mapping system against precipitation data gathered and archived at the ARS research laboratory over several decades. For the central Oklahoma region, the PRISM monthly precipitation climatology was shown to lack accuracy for the site under consideration, underestimating both average and maximum monthly rainfall. Even though these findings are preliminary and for one location, the observed lack of accuracy is of particular concern, since it was expected that the PRISM estimates would be relatively accurate in this non-mountainous area. If subsequent expanded investigations confirm this result, then the PRISM data product should be used with extreme caution when developing climate-informed decision support for agriculture.