Monitoring and comparing NDSI changes using MODIS and ETM+ sensor data to estimate snow cover in North Karun Basin

North Karun Basin is one of the snow-covered regions of Iran and its snowfall has a great role in the water supply situation in the central and southern regions of the country. In this research, an attempt has been made to investigate the changes in the snow-covered surface in the study basin. Access to this information in high and impassable areas is possible only with the help of remote sensing techniques. The MODIS sensor has multiple spectral bands, but the ETM+ sensor has a better spatial resolution. The purpose of this study is to monitor, compare and map snow-covered areas by MODIS and ETM+ data in the northern part of the Karun Basin.

The research has been conducted in the northern part of Karun Basin, as one of the snowy areas in Iran. This area is geographically between 31° 19' 9'' to 31° 39' 17'' northern latitude and 49° 33' 56'' to 51° 54' 23'' eastern longitude. The study basin is over 14802 square kilometers, about 90% of Chaharmahal and Bakhtiari Province and 23% of Karun Basin. The height of the study area is changing between 800 m at Karun No. 4 dam to 4440 m at the Peak of Zard Kooh Mountain. The satellite data included MODIS and ETM+ sensors over a period of 15 years (2000 to 2015) used to calculate NDSI. The satellite pass-times were at Julian day number 360, 333, 365, 355, 360, 363, 331, 350, 336, 355, 342, 345, 363, 334 and 1.

The minimum and maximum snow-covered areas produced by MODIS in 2010 and 2013 were 107,295 and 1,364,118 ha, respectively, and for ETM+ sensor, were 128,758 and 1,090,580 ha, in 2010 and 2006, respectively. The snow-covered areas estimated by MODIS in 14 dates were more than the corresponding values for ETM+. The exception was 2011 when the snow-covered area in the ETM+ was greater than MODIS (3067 ha or 1.36%). Based on the results in 15 years, the MODIS sensor overestimated the snow-covered area equivalent to 26.95% compared to ETM+. In a small snow-covered area, deviation from the 1:1 line was small but, by increasing snow-covered area, deviation increased. The overall difference in snow pixels in the 15 years between the two sensors was 26.95%. As MODIS pixels are 500 by 500 m, and ETM+ pixels are 30 by 30 m, the size of the pixels in ETM+ resized to 500 m by resampling method in which to use same pixel size for another comparison. After resampling, the slope of the regression line slightly dropped. However, there are no-significant differences. Using data from MODIS and ETM+ sensors had a good performance to monitor spatio-temporal and map snow-covered surfaces in North Karun Basin.

Due to spatio-temporal frequency in obtaining MODIS imageries via internet freely in temporal resolution of eight-days and one-day, and also at very wide area coverage, short-time monitoring is possible for hydrological studies, forecasting and flood warning. On the other hand, the ability and better spatial resolution of ETM+, small snowy fields could be estimated and mapped better than MODIS. Although it may be need to mosaic several ETM+ imageries to cover wide areas. However, paying attention to geographical characteristics and time of the study area, end user decides to choose one or both sensor data.