Tuesday, November 14, 2006

Evaluation of Geospatial Models for Estimating Nutrient Distribution in a Vegetative Treatment Area Used or Feedlot Runoff Control.

Bryan Woodbury, Roger Eigenberg, and Jack Nienaber. PO Box 166, State Spur 18d, Clay Center, NE 68933

Liquid runoff from cattle feeding operations can be: 1) held in long-term storage or 2) temporarily held in a settling basin, then discharged to a vegetative treatment area (VTA) for nutrient utilization.  The second option has been shown to be effective on the short-term, but long term sustainability is yet to be determined.  Understanding the spatial nutrient distribution within the VTA would allow for total nutrient mass to be monitored year to year for evaluating long-term sustainability.  Geostatistical estimation methods can involve the nutrient of interest as a primary variate combined with a lower cost secondary variate such as soil conductivity.  Prediction methods that use a primary and secondary variate often rely on cokriging as the method for spatial nutrient distribution estimation.  Previous work with cokriging has shown it to provide only marginal improvement over simple kriging methods.  Also, cokriging requires a large number of expensive samples to adequately estimate the primary variate spatial distribution. This study was conducted to compare estimation methods for determining spatial distribution of nutrients across a VTA.  Multiple linear regression methods will be investigated in addition to simple kriging, cokriging.  A research VTA for a cattle feedlot at the U.S. Meat Animal Research Center (USMARC) at Clay Center, NE serves as the test facility for the study.