Understanding irrigator responses to previous changes in water availability is critical to building effective institutions that allow for efficient and resilient management of water resources in the face of potentially increasing scarcity due to climate change. Using remote sensing data, I examined irrigator responses to seasonal changes in water availability in Idaho's Snake River Plain over the past 33 years. Google Earth Engine's high performance cloud computing and big data processing capabilities were used to compare the performance of three spectral indices, three compositing algorithms and two sensors for 2002 and 2007 for distinguishing between irrigated and non-irrigated parcels. We demonstrate that, on average, the seasonal-maximum algorithm yields a 60% reduction in county scale root mean square error (RMSE) over the accepted single-date approach. We use the best performing classification method, a binary threshold of the seasonal maximum of the Normalized Difference Moisture Index (NDMI), to identify irrigated and non-irrigated lands in Idaho's Snake River Basin for 1984-2016 using Landsat 5-8 data. NDMI of irrigated lands was found to generally increase over time, likely as a result of changes in agricultural practices increasing crop productivity. Furthermore, we find that irrigators with rights to small areas, and those with only surface water rights are more likely to have a major reduction (>25%) in irrigated area and conversely those with a large, groundwater rights are more likely to have major increases (>25%) in the extent of their irrigation. / Master of Science / Understanding irrigator responses to previous changes in water availability is critical to building effective institutions that allow for efficient and resilient management of water resources in the face of potentially increasing scarcity due to climate change. Using remote sensing data, I examined irrigator responses to seasonal changes in water availability in Idaho’s Snake River Plain over the past 33 years. Google Earth Engine’s high performance cloud computing and big data processing capabilities were used to compare the performance of three spectral indices, three compositing algorithms and two sensors for 2002 and 2007 for distinguishing between irrigated and non-irrigated parcels. We demonstrate that, on average, the seasonal-maximum algorithm yields a 60% reduction in county scale root mean square error (RMSE) over the accepted single-date approach. We use the best performing classification method, a binary threshold of the seasonal maximum of the Normalized Difference Moisture Index (NDMI), to identify irrigated and non-irrigated lands in Idaho’s Snake River Basin for 1984-2016 using Landsat 5-8 data. NDMI of irrigated lands was found to generally increase over time, likely as a result of changes in agricultural practices increasing crop productivity. Furthermore, we find that irrigators with rights to small areas, and those with only surface water rights are more likely to have a major reduction (>25%) in irrigated area and conversely those with a large, groundwater rights are more likely to have major increases (>25%) in the extent of their irrigation.
| Identifer | oai:union.ndltd.org:VTETD/oai:vtechworks.lib.vt.edu:10919/78361 |
| Date | 18 July 2017 |
| Creators | Chance, Eric Wilson |
| Contributors | Forest Resources and Environmental Conservation, Cobourn, Kelly M., Thomas, Valerie A., McGuire, Kevin J., Wynne, Randolph H. |
| Publisher | Virginia Tech |
| Source Sets | Virginia Tech Theses and Dissertation |
| Detected Language | English |
| Type | Thesis |
| Format | ETD, application/pdf |
| Rights | In Copyright, http://rightsstatements.org/vocab/InC/1.0/ |
Page generated in 0.236 seconds