Estimation of wet snow cover area with SAR imagery in the basin of Tupungato River, Mendoza, Argentina
| Autor: | Sofia A. Teverovsky, Claudia Notarnicola, Graciela Salinas de Salmuni, Ignacio G. Pascual |
|---|---|
| Rok vydání: | 2018 |
| Předmět: |
geography
geography.geographical_feature_category 010504 meteorology & atmospheric sciences 0211 other engineering and technologies Drainage basin 02 engineering and technology Structural basin Snow 01 natural sciences Dry snow Geomorphology Geology Snow cover 021101 geological & geomatics engineering 0105 earth and related environmental sciences |
| Zdroj: | 2018 IEEE Biennial Congress of Argentina (ARGENCON). |
| Popis: | The data coming from remote sensing, in particular Synthetic Aperture Radars (SAR), have great potential in the study of snow since they provide information regardless of weather conditions. In this work, the capacity of the C band in wet snow detection is investigated, in the Tupungato river basin, province of Mendoza, Argentina. To this scope, Sentinel 1 images were used, processed following the methodology developed by Nagler and Rott (2000), which adopts a change detection technique by considering an image taken as a reference, in a snow-free or dry snow condition. This algorithm requires the adaptation of parameters according to the particular characteristics of the study area. The analysis carried out in this basin indicates that the optimal threshold to identify wet snow is -2dB. The results were validated indirectly by using information of surface temperature and snow cover area, obtained with LANDSAT 8 optical images. In this way, it was verified the correct classification with SAR of pixels corresponding to wet snow. Wet snow maps generated with SAR data are very useful to nourish hydrological models for the forecast of flow in mountain areas dominated by snow regime.Los datos provenientes de sensores remotos, en particular de Radares de Apertura Sintetica (SAR), poseen gran potencial en el estudio de la nieve ya que proveen informacion independientemente de la condicion atmosferica reinante. En este trabajo se investiga la capacidad de la banda C en la deteccion de nieve humeda, en la cuenca del rio Tupungato, provincia de Mendoza, Argentina. Para ello se utilizaron imagenes Sentinel 1, procesadas siguiendo la metodologia desarrollada por Nagler y Rott (2000) que emplea la tecnica de deteccion de cambios respecto a una imagen tomada como referencia, en condicion libre de nieve o con nieve seca. Este algoritmo requiere la adecuacion de parametros en funcion a las caracteristicas particulares del area de estudio. El analisis realizado en la cuenca indico que el umbral optimo para identificar coberturas de nieve humeda es -2dB. Los resultados fueron validados indirectamente a partir de datos de temperatura de superficie y area cubierta de nieve, obtenidos con imagenes opticas LANDSAT 8. De esta forma, se verifico la correcta clasificacion con SAR de pixeles correspondientes a nieve humeda. Los mapas de nieve humeda generados con datos SAR son de mucha utilidad ya que nutren a modelos hidrologicos para el pronostico de caudal en zonas con regimen nival. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|