Irrigated Acreage Identification and Assesment Final Report

Home Browse Search

• Division 4's attribute data needs to be refined. In 1993, the thought process was to tie structures to the field they irrigate. Often this can be complex. In an attempt to more accurately describe the situation, Division 4 assigned 2 or more structures to a field, and gave each a percentage. This has since been difficult to use. It is recommended that Division 4 use the "1 Field — 1 Structure ID" method as the other divisions did. There will be a reduction in accuracy but it will be standard with the other divisions and easier to use. The pilot study did not fully resolve this issue and the Division of Water Resources needs to investigate further. C. Suggested Methodology for completing Irrigated Acreage Pilot Study Step 1 • Define Area of Interest (AOI), and prepare the data. o Define the AOI o Obtain the imagery o Clip and cut the images, shapefiles, and coverages to the AOI o Make sure all projections are the same: UTM Zone 13, NAD27, Meters, Clarke1866 Step 2 • Define Polygons using DOQQ's o Display the 1993 polygon coverage, overlaid on the DOQQ's. Use both together to guide your digitizing of the polygons. o Visually digitize each polygon. When a polygon is reshaped, split, or added note the edit in the attribute table. Step 3 • Classify imagery, and identify crop types. o Use the Principle component analysis (transformation analog) tools to compress the 7 bands of imagery into 3 bands. o Perform an unsupervised classification. In the pilot project 50 classes were defined, with a .970 convergence, and 15 iterations. o Use the unsupervised classification to guide your supervised classification. We used 50 unsupervised classes plus 62 supervised classes. The final project was 17 —23 different crop types. Step 4 • Use Tassel Cap process to identify irrigated vs. non-irrigated lands. o The Tassel Cap approach highlights data from bands 2 (wetness), and 3 (greenness). From this image, an unsupervised classification identified 55 classes, which were combined into 3 classes; water, irrigated, and not irrigated. Step 5 • Move crop type, irrigation status to the new 1999 map. o Once the classification was completed, the image was converted into a vector map by crop type. Using Erdas and Arc/Info AMLs the imagery and new polygon maps were overlaid, and the attributes transferred from the imagery to the new polygons. See AML's in Exhibit D. 4