Artificial neural network coupled with wavelet transform for estimating snow water equivalent using passive microwave data

Abstract

Snow Water Equivalent (SWE) is an important parameter in hydrologic engineering involving the streamflow forecasting of high-elevation watersheds. In this paper, the application of classic Artificial Neural Network model (ANN) and a hybrid model combining the wavelet and ANN (WANN) is investigated in estimating the value of SWE in a mountainous basin. In addition, k-fold cross validation method is used in order to achieve a more reliable and robust model. In this regard, microwave images acquired from Spectral Sensor Microwave Imager (SSM/I) are used to estimate the SWE of Tehran sub-basins during 1992-2008 period. Also for obtaining measured SWE within the corresponding Equal-Area Scalable Earth-Grid (EASE-Grid) cell of SSM/I image, approach of Cell-SWE extraction using height-SWE relations is applied in order to reach more precise estimations. The obtained results reveal that the wavelet-ANN model significantly increases the accuracy of estimations, mainly because of using multiscale time series as the ANN inputs. The Nash-Sutcliffe Index (NSE) for ANN and WANN models is respectively 0.09 and 0.44 which shows a firm improvement of 0.35 in NSE parameter when WANN is applied. Similar trend is observed in other parameters including RMSE where the value is 0.3 for ANN and 0.07 for WANN.

Snow depth retrieval from microwave remote sensing by combining wavelet transform and machine learning models in Northern Xinjiang, China

“…Nonetheless, they may struggle to extract information from nonstationary signals and can easily get trapped in local optima. 1,34 To mitigate these challenges, researchers integrated wavelet transform with ANN models to overcome the impacts of internal snow properties and external environmental factors, such as feedforward neural networks (FFNN) and general regression neural networks (GRNN), with wavelet transform. 34,35 The wavelet transform technique has been shown to enhance the capabilities of ANN models in estimating the snow depth in regions with complex topography and variable snow cover.…”

“…1,34 To mitigate these challenges, researchers integrated wavelet transform with ANN models to overcome the impacts of internal snow properties and external environmental factors, such as feedforward neural networks (FFNN) and general regression neural networks (GRNN), with wavelet transform. 34,35 The wavelet transform technique has been shown to enhance the capabilities of ANN models in estimating the snow depth in regions with complex topography and variable snow cover. [36][37][38] Adib et al 39 advanced snow depth retrieval by integrating multiple artificial intelligence models and a multiresolution-based maximal overlap discrete wavelet transform (MODWT-MRA).…”

“…The above studies indicate that the ANN can improve the estimation accuracy of snow products, especially in mountainous areas. Nonetheless, they may struggle to extract information from nonstationary signals and can easily get trapped in local optima 1 , 34 …”

Spatiotemporal estimation of snow depth using point data from snow stakes, digital terrain models, and satellite data

“…Many studies have been developed to map SWE for dates with available ground‐truth data, matching them with satellite‐based observations (Harshburger et al, ) applying multiple regression techniques). Dariane, Azimi, and Zakerinejad () applied an ANN coupled with wavelet transform to estimate the SWE using passive microwave data.…”