Laserfiche WebLink
SPDSS Memorandum <br />To: Ray Alvarado, Ray Bennett <br />From: Riverside Technology, inc., Timothy Martin and Claudio Schneider <br />Subject: SPDSS Draft Memorandum for Task 89.1 -Finalize Methods and Order Imagery <br />Date: 8 July 2003 <br />1. Introduction <br />This Spatial Systems Integration component of the South Platte Decision Support System (SPDSS) is <br />mapping and classifying current and historic land use, irrigated lands and irrigation service areas for both <br />ground and surface water. The purpose of this memo is to review relevant literature, including reports of <br />researchers and previous State contractors engaged in mapping land use and crop types for water <br />resources applications, and consequently describe the recommended technical approach for mapping <br />current land use. This memo also recommends the full list of Landsat satellite images for purchase and <br />use in the SPDSS. This document compliments the SPDSS Feasibility Study (2000) and the Scope of <br />Work for the Spatial System Integration (October, 2002) with background and more detail on the <br />approach for mapping irrigated lands and crop types. <br />The use of optical, multi-spectral satellite imagery (e.g., Landsat TM data) has significant advantages <br />over traditional photo-interpretation methods when mapping large areas (Jensen 1996, Lillesand and <br />Kiefer 2000). Among the advantages of satellite imagery are: the reduced time and cost needed to <br />produce classifications; fast and inexpensive updating due to the temporal resolution of the satellites; <br />additional information provided by the multi-spectral images; and economies of scale with digital image <br />processing for classification and cumulative effect or change analysis. <br />2. Remote sensing application for irrigated lands assessment <br />In recent years, several satellite classification approaches have been conducted in Colorado and other <br />regions to classify agricultural crops and irrigated lands. These studies provide a background for <br />developing classification procedures for mapping current land use and irrigated field boundaries of the <br />South Platte Decision Support System Project (SPDSS). A summary of the studies and projects is <br />presented in Table 1. <br />2.1 Previous classification of irrigated lands by the State of Colorado <br />Several approaches have been used in recent classifications of irrigated lands in the State of Colorado. <br />Agro Engineering (2000) manually digitized irrigated field boundaries in the Irrigated Lands Assessment <br />Using Satellite Imagery in the Rio Grande Basin of Colorado after failed attempts to use automated edge <br />detection techniques. Irrigated parcels were mapped using Landsat and IRS-1C satellite imagery in <br />conjunction with other available data. A total of 625,364 acres was classified as irrigated in 1998. The <br />report stated that this estimate of total irrigated acreage for major harvestable crops agreed within 91 <br />percent of Colorado Agricultural Statistics for 1998. The accuracy of the overall classification was <br />evaluated using ground truth data collected by the Rio Grande Water Conservation District. The overall <br />accuracy obtained in this evaluation was reported as 92 percent. The techniques used to compute this <br />overall accuracy assessment are not conventional and no accuracy of results was reported for individual <br />crops. <br />Page 1 of 13 ~-~~versfde TecAnafagy, lnc. <br />W:rler Resources Frr~mec.•rng ,end L'nr+surlrn2 <br />