Does the U.S. Cropland Data Layer Provide an Accurate Benchmark for Land‐Use Change Estimates?

Autor: Kurtis D. Reitsma, Sharon A. Clay, David E. Clay, Cheryl Reese, Barry H. Dunn
Rok vydání: 2016
Předmět:
Zdroj: Agronomy Journal. 108:266-272
ISSN: 1435-0645
0002-1962
DOI: 10.2134/agronj2015.0288
Popis: Agronomy Journa l • Volume 108 , I s sue 1 • 2016 The U.S. CDL is an annual raster-based land-use map created by the USDA, National Agricultural Statistic Service (NASS). Th e CDL predicted land-uses are based on the refl ective signatures from a number of satellites including Landsat 5 TM, Landsat 7 ETM+, the Indian Remote Sensing RESOURCESAT-1 (IRS-P6), Advanced Wide Field Sensors (AWiFS), Disaster Monitoring Constellation (DMC) DEIMOS-1, and UK2 sensors (Edinger, 2012). Th e CDL is being used for many diff erent purposes ranging from making disaster assessments to making agricultural policy decisions (Fernandez-Cornejo and Caswell, 2006; Maitima et al., 2009; Chang et al., 2007; Carpenter, 2010; Hatfi eld et al., 2011; Schrag, 2011; Han et al., 2012; Li et al., 2012; Bandaru et al., 2013; Wright and Wimberly, 2013a; Decision Innovation Solutions, 2013; Johnston, 2013; Johnson, 2013; Mueller and Harris, 2013; Clay et al., 2014; IPCC, 2014; Elliot et al., 2014; Lee et al., 2014; Liska et al., 2014; Reitsma et al., 2015). Th e CDL can be accessed through CropScape, which allows the user to query the database for specifi c information (Han et al., 2012; http://nassgeodata.gmu.edu/CropScape/). Statewide CDL meta data for each year is posted for crops that include corn (Zea mays L.), pea (Pisium sativum L.), and alfalfa (Medicago sativa L.). Each crop has a diff erent user (percent of CDL classifi ed points that were correctly characterized) and producer (percent of ground collected sites that were correctly identifi ed) accuracies. For example, in 2006 corn and alfalfa grown in South Dakota had producer accuracies of 74.63 and 27.63%, respectively. However, meta data for many land uses including grass/pasture, mixed forest, and shrubland were not posted. Th e CDL provides digital data that can be used to determine land-use changes. Unfortunately, diff erent processing approaches can produce diff erent answers (Decision Innovation Solutions, 2013; Wright and Wimberly, 2013a, 2013b; Kline et al., 2013; Laingen, 2015). For example, Wright and Wimberly (2013) reported that from 2006 to 2011 there was a net loss of 182,000 ha of grassland in South Dakota, whereas Decision Innovation Solutions (2013) reported that for the same area there was grassland decline of 879,000 ha from 2007 to 2012. Resolving diff erences between these studies are important because data derived or processed from the CDL is being widely distributed and used to infl uence public policies. For example, is crop insurance contributing to the plowing of native prairies, and is ethanol production contributing to the declines in wildlife populations. Th erefore, Crop Economics, Production & Management
Databáze: OpenAIRE
Externí odkaz: