Application of satellite-borne microwaves in estimation of snowpack properties.
Item
-
Title
-
Application of satellite-borne microwaves in estimation of snowpack properties.
-
Identifier
-
AAI3249910
-
identifier
-
3249910
-
Creator
-
Eshraghi Azar, Amir.
-
Contributor
-
Advisers: Reza Khanbilvardi | Hosni Ghedira
-
Date
-
2007
-
Language
-
English
-
Publisher
-
City University of New York.
-
Subject
-
Engineering, Civil | Environmental Sciences | Hydrology | Remote Sensing
-
Abstract
-
This study investigates different approaches and proposes a new algorithm to estimate Snow Water Equivalent (SWE) in the Great Lakes area using satellite microwave data. The research focuses on implementing land cover variation into SWE estimating models in order to improve the estimation of SWE.;The study area is located between latitudes 41┬░N-49┬░N and longitudes 87┬░W-98┬░W on the transitional zone of snow with thaw and refreeze events during the snow season. Additionally, the study includes a variety of land cover types including evergreen needle leaf forest, deciduous broadleaf forest, cropland, woodland and dry land. Various approaches were tested in the study area for SWE estimations. These approaches are divided into: (1) Using high resolution active microwave RADARSAT SAR to estimate SWE. This section focuses on how to process RADARSAT images and corresponding ground truth data, and suggests modeling approaches to improve the estimations. The results revealed low correlations between estimated and ground truth SWE; however, adding NDVI data in order to introduce land cover variation improved results more than 15 percent. (2) Using low resolution passive SSM/I and active QuikSCAT-Ku with statistical based models. The developed algorithms were first tested to identify snow cover. Then, the model was tested to estimate SWE. The improvement by combining various data types in an Artificial Neural Network (ANN) model was quantified. The ANN model shows satisfactory results in dependent estimations of snow cover. In SWE estimation NDVI improved the accuracy of the results. (3) Using a physical based model to estimate SWE. A non-linear physical based model was developed and tested in the study area. The algorithm was developed based on the analysis of SSM/I scattering signatures and NDVI data with SWE and snow depth data. The results showed that the NDVI value is directly correlated with factors of linear algorithms. Using SSM/I along with NDVI data, a non-linear algorithm was proposed to estimate SWE. The temporal validation of the proposed algorithm for different test sites showed more than 20 improvements in Root Mean Square Error (RMSE). The spatial validation for the whole study area illustrated consistent improvements in both correlation coefficients and RMSE. Specifically, in forested areas the RMSE decreased approximately 50 percent indicating a significant improvement in the estimations. In summary, the proposed algorithm improves the estimation of snow depth and Snow Water Equivalent (SWE) about 30 percent.
-
Type
-
dissertation
-
Source
-
PQT Legacy Restricted.xlsx
-
degree
-
Ph.D.