111-9
Effect of Sample Size On the Accuracy of Geospatial Mapping.

Poster Number 506

Monday, November 4, 2013
Tampa Convention Center, East Hall, Third Floor

Chase M Straw1, Gerald M. Henry2, Robert N. Carrow3, Van W. Cline4, Kevin Tucker5 and Robin Landry5, (1)University of Georgia-Athens, Athens, GA
(2)Crop and Soil Sciences, University of Georgia-Athens, Athens, GA
(3)University of Georgia - Griffin, Griffin, GA
(4)Toro Company, Bloomington, MN
(5)Crop and Soil Sciences, University of Georgia, Athens, GA
Global Information System (GIS) technology allows for interpolation of collected data points to predict continuous values across an area. The Toro Precision Sense 6000 is a mobile device, used for turfgrass systems, that simultaneously collects and geo-references percent volumetric water content (%VWC), soil penetration resistance, normalized difference vegetation index (NDVI), soil salinity, and elevation characteristics at exact locations across an area. For an athletic field, using a 2.4 m x 2.4 m sample grid, can generate over 1000 data samples that can be collected and interpolated accurately assess the performance of that particular field. Previous assessment methods for athletic field performance testing (American Society for Testing and Material) employed hand-held sensors to collect data from only 6-8 locations across a field. This small sample size may make it very difficult to completely and accurately evaluate the entire field. Although increasing the number of samples is necessary to enhance precision, little research has been conducted to investigate map accuracy in response to sample size. The introduction of the Precision Sense 6000 allows for quick and intense data collection with the ability to manipulate sample spacing and overall number. Surveys were conducted on three soccer fields at a sports complex in Roswell, GA and another in Athens, GA during the summer of 2013. Each sports complex was considered an experimental run, whereas soccer fields were considered replications within each run. Four sample sizes of volumetric water content (soil moisture) data were obtained using the Toro Precision Sense 6000 unit and interpolated/evaluated using ArcGIS software in order to determine the accuracy of each sample size to predict unknown values across each entire field. The sample sizes evaluated consisted of 1068 (2.4 m x 2.4 m sample grid), 535 (2.4 m x 4.9 m sample grid), 141 (4.9 m x 9.8 m sample grid), and 24 (19.5 m x 19.5 m sample grid). Cross validation was performed to determine the effect of sample size on the accuracy of each prediction model. Results from this research can be used to determine a standard method for data collection using the Toro Precision Sense 6000.
See more from this Division: C05 Turfgrass Science
See more from this Session: Turfgrass Cultural Practices and Weed Control

Show comments