Snowmelt Flood Mapping and Land Surface Short-Term Dynamics Assessment in a “Before-During-After” Scenario Based on Radar and Optical Satellite Imagery: Case Study Around the Lewisville Lake (Dallas/Fort Worth Metropolitan, Texas, USA)  

作　　者：Alfred Homère Ngandam Mfondoum Roseline Batcha Alfred Homère Ngandam Mfondoum;Roseline Batcha(Stats’NMaps, Private Consulting Firm, Dallas, USA;Department of Geography, University of Yaoundé I, Yaoundé, Cameroon)

机构地区：[1]Stats’NMaps, Private Consulting Firm, Dallas, USA [2]Department of Geography, University of Yaoundé I, Yaoundé, Cameroon

出　　处：《Advances in Remote Sensing》2023年第1期1-28,共28页遥感技术进展（英文）

摘　　要：The main goal of this study has been to map flood and assess land surface short-term dynamics in relation with snowy weather. The two recent snowfall events, which happened in, February 14^th and 15^th, of year 2021, and February 3^rd and 4^th, of year 2022, were chosen. A pre-analysis correlation was assumed between, the snow events, recurrency of floods, and changes in the land surface characteristics (i.e., wetness, energy, temperature), in a “Before-During-After” scenario. Active and passive microwave satellites data such as, Sentinel-1 synthetic aperture radar (SAR), Sentinel-2 multispectral instrument (MSI) and Landsat-9 Operation Land Imager-2/Thermal Infrared Sensors-2 (OLI-2/TIRS-2), as well as cloud databased global models for water and urban layers were used. The first step of processing was thresholding of SAR image, at 0.25 cutoff, based on bimodal histogram distribution, followed by the change analysis. The following processing consisted in the images transformation, by computing the tasseled cap transformation wetness (TCTw) and the surface albedo on MSI image. In addition, the land surface temperature (LST) was modeled from OLI-2/TIRS-2 image. Then, a 5^th order polynomial regression was computed, between TCTw as dependent variable and, albedo and LST as independent variables. As a first result, an area of 5.6 km² has been mapped as recurrently flooded from the two years assessment. The other output highlighted a constant increase of wetness (TCTw), considered most influential on land surface dynamics, comparatively to energy exchange (albedo) and temperature (LST). The “After” event dependency between the three indicators was highest, with a correlation coefficient, R² = 0.682, confirming the persistence of wetness after-snowmelt. Validation over topographic layers confirmed that, recurrently flooded areas are mostly distributed on, lowest valley depth points, farthest distances from channel network (i.e., from perennThe main goal of this study has been to map flood and assess land surface short-term dynamics in relation with snowy weather. The two recent snowfall events, which happened in, February 14^th and 15^th, of year 2021, and February 3^rd and 4^th, of year 2022, were chosen. A pre-analysis correlation was assumed between, the snow events, recurrency of floods, and changes in the land surface characteristics (i.e., wetness, energy, temperature), in a “Before-During-After” scenario. Active and passive microwave satellites data such as, Sentinel-1 synthetic aperture radar (SAR), Sentinel-2 multispectral instrument (MSI) and Landsat-9 Operation Land Imager-2/Thermal Infrared Sensors-2 (OLI-2/TIRS-2), as well as cloud databased global models for water and urban layers were used. The first step of processing was thresholding of SAR image, at 0.25 cutoff, based on bimodal histogram distribution, followed by the change analysis. The following processing consisted in the images transformation, by computing the tasseled cap transformation wetness (TCTw) and the surface albedo on MSI image. In addition, the land surface temperature (LST) was modeled from OLI-2/TIRS-2 image. Then, a 5^th order polynomial regression was computed, between TCTw as dependent variable and, albedo and LST as independent variables. As a first result, an area of 5.6 km² has been mapped as recurrently flooded from the two years assessment. The other output highlighted a constant increase of wetness (TCTw), considered most influential on land surface dynamics, comparatively to energy exchange (albedo) and temperature (LST). The “After” event dependency between the three indicators was highest, with a correlation coefficient, R² = 0.682, confirming the persistence of wetness after-snowmelt. Validation over topographic layers confirmed that, recurrently flooded areas are mostly distributed on, lowest valley depth points, farthest distances from channel network (i.e., from perenn

关 键 词：Snowmelt Flood Short-Term Dynamics RADAR Optical Lewisville Lake 

分 类 号：TN9[电子电信—信息与通信工程]