Assessment of ice mapping system and moderate resolution imaging spectroradiometer snow cover maps over Colorado Plateau

Abstract

Satellite snow cover area (SCA) mapping using optical sensors has been known to suffer severe obstruction due to vegetation canopy and cloud cover. Several algorithms have been developed to reduce cloud cover contamination and enhance the SCA mapping. In this study we introduce the use of a daily SCA product from the Multisensor Snow and Ice Mapping System (IMS) at a nominal resolution of 4 km, assess its accuracy and error levels against in situ observations, and compare the IMS SCA product with the SCA products from moderate resolution imaging spectroradiometer (MODIS), a combination of daily Terra and Aqua satellites. The results show that the snow accuracies are higher during winter for both IMS and MODIS, and that there is not much difference between MODIS at 500 m and upscaled at 4 km. The IMS SCA mapping accuracies are significantly higher than MODIS accuracies for all sky conditions, while they are similar to or slightly lower than MODIS accuracies in clear sky conditions. The overestimate error of snow cover using IMS is higher (lower) than that of MODIS during ablation (accumulation) periods. Both MODIS and IMS show a similar pattern of underestimation errors of snow cover with the IMS being slightly higher than the MODIS. It is concluded that the IMS SCA product has potential as a good alternative for the MODIS daily SCA products or replacing those cloud pixels in the MODIS daily or multiday products.

Tracking Snow Variations in the Northern Hemisphere Using Multi-Source Remote Sensing Data (2000–2015)

“…Although the combination of MODIS and AMSR-E/IMS can take advantage of both high spatial resolution of optical data and cloud transparency of passive microwave data, the combination might reduce the spatial resolution of MODIS products because cloud pixels in MODIS are replaced by AMSR-E/IMS images, which have coarse spatial resolution; therefore, the accuracy of the cloud pixel reclassification fully depends on the accuracy of AMSR-E and IMS. In previous studies, AMSR-E and IMS data were widely used for monitoring snow, and both of them have relatively high accuracy [28][29][30][31]. On the other hand, because of the poor performance in the forest area, the Kappa test of integrated snow products shows moderate consistency with Landsat, and snow monitoring in this regions also remains a major problem for virtually all types of global remote sensing snow products, for which overall accuracy were seriously affected due to the forest canopy [37].…”

Developing Daily Cloud‐Free Snow Composite Products From MODIS and IMS for the Tienshan Mountains

“…The overall accuracy (Oc, Oa), snow accuracy (Sc, Sa), and land accuracy (Lc, La) indices (Mazari et al, 2013) are calculated for clear-and all-sky conditions based on the confusion matrix between satellite data versus station measurement. The two error types evaluated are image underestimate error (IU) and image overestimate error (IO; Gao et al, 2010;Parajka & BlöSchl, 2008).…”

“…For example, Yu et al (2016) used this method to realize cloud-free snow products on the Tibetan Plateau and the overall accuracy of the new product is 94% as compared with ground stations in allsky conditions. But it is worth mentioning that Mazari et al (2013) indicated that the overestimate error of the IMS is 5% more than that of the original MODIS during the ablation period and 1% to 2% less than original MODIS during the accumulation period. Therefore, to further reduce the possible errors from IMS, we first use temporal and spatial filtering to remove some clouds before combining MODIS and IMS.…”

“…As the water tower of Central Asia, the Tienshan Mountain's glacier and snow cover melt water supplies the major rivers in the area (e.g., Syr Darya River, Ili River, and Tarim River) (Chen et al, 2016). Given the importance of these water bodies to agriculture and animal husbandry in the downstream arid regions (Kong & Pang, 2012), accurate information on snow cover is essential for understanding details of climate change and water management (Brown & Robinson, 2010;Mazari et al, 2013).…”

Use of interactive multisensor snow and ice mapping system snow cover maps (IMS) and artificial neural networks for simulating river discharges in Eastern Turkey

“…Specifically, during the learning phase, connection weights are changed continuously in an adaptive manner, so that the mean square error between the observed and predicted output values is minimum. Extensive descriptions about working principles of ANN can be accessed fromLohani et al (2012).Runoff values observed in UER were simulated byTekeli et al (2005a) using daily average temperature, daily cumulated precipitation, and SCA obtained from MODIS sensor as input variables into snowmelt runoff model (SRM) for water year 2004. Also,…”