Seasonal snow and ice support important ecosystems and comprise the main water supply for nearly two billion people, yet forecasts of snowmelt runoff either do not exist or occasionally significantly under- or over-predict flows. Mapping snow cover and albedo from space has enabled a paradigm shift in our ability to observe processes that occur in the mountains. However, clouds commonly obstruct the surface in visible and infrared spectra, and existing operational cloud masks consistently misclassify some snow as clouds and some clouds as snow. This work aims to both advance our ability to utilize satellites for observing the mountain snowpack and to improve our understanding of the potential operational benefits of better forecasts from these tools.
In Chapter 1, I assess the errors in the cloud masks over snow-covered, mid-latitude mountains for the Landsat 8 Operational Line Imager (OLI) and for the Moderate-Resolution Imaging Spectroradiometer (MODIS) on the Terra satellite: CFMask for Landsat 8 and the cloud mask that ships with the MOD09GA and MYD09GA surface reflectance products from MODIS. The overall precision and recall of CFMask over snow-covered terrain are 0.70 and 0.86; the MOD09GA cloud mask precision and recall are 0.17 and 0.72. I find a plausible reason for poorer performance of cloud masks over snow lies in the potential similarity between multispectral signatures of snow and cloud pixels in three situations: (1) Snow at high elevation is bright enough in the “cirrus” bands (Landsat band 9 or MODIS band 26) to be classified as cirrus. (2) Reflectances of “dark” clouds in shortwave infrared (SWIR) bands are bracketed by snow spectra in those wavelengths. (3) Snow as part of a fractional mixture in a pixel with soils sometimes produces “bright SWIR” pixels that look like clouds.
In Chapter 2, I develop a new method for snow/cloud discrimination that relies on textural and spatial features alongside the spectra to identify clouds and their optical properties. I extend the applicability of superpixels to multispectral satellite data by generating superpixels from an 8-D space of the eight 30m Landsat 8 optical bands by mapping the spectral angle between superpixel mean spectra and each component pixel. A Gabor filter bank on Landsat 8 panchromatic data accurately discriminates mountain snow that is spectrally similar to clouds. Clouds and shadows are also identified by using the expected topographic shading alongside spectral tests to separate clouds, terrain shadow, cloud shadow, and shaded snow. Histograms of oriented gradients identify mountainous terrain and obstruction of terrain illumination patterns by clouds. A radiative transfer model assigns optical properties to cloud mask pixels. This new method improves the precision of Landsat 8 cloud masks by 31% (from 70% to 91%) over mid-latitude snow covered mountains compared to the operational product.
To understand the opportunities for remote sensing of snow for water operations, in Chapter 3 I use thirty-four years of water data in fourteen California Sierra Nevada basins to examine the association between water management decisions and snowmelt runoff forecasts, their uncertainty, and error. The findings indicate that forecasts are positively associated with releases, but increased uncertainty in a forecast is negatively associated with releases. Reduction in spring runoff forecast uncertainty is possible from remote sensing tools and would enable additional uses of storing runoff from April through July.
|Commitee:||Roberts, Dar, Costello, Chris, Bales, Roger|
|School:||University of California, Santa Barbara|
|Department:||Environmental Science & Management|
|School Location:||United States -- California|
|Source:||DAI-B 81/8(E), Dissertation Abstracts International|
|Subjects:||Remote sensing, Water Resources Management, Physical geography|
|Keywords:||Cloud, Forecast, Radiative transfer, Snow, Texture|
Copyright in each Dissertation and Thesis is retained by the author. All Rights Reserved
The supplemental file or files you are about to download were provided to ProQuest by the author as part of a
dissertation or thesis. The supplemental files are provided "AS IS" without warranty. ProQuest is not responsible for the
content, format or impact on the supplemental file(s) on our system. in some cases, the file type may be unknown or
may be a .exe file. We recommend caution as you open such files.
Copyright of the original materials contained in the supplemental file is retained by the author and your access to the
supplemental files is subject to the ProQuest Terms and Conditions of use.
Depending on the size of the file(s) you are downloading, the system may take some time to download them. Please be