Satellite-Based Rainfall Retrieval: From Generalized Linear Models to Artificial Neural Networks

Abstract

Abstract:In this study, we develop and compare satellite rainfall retrievals based on generalized linear models and artificial neural networks. Both approaches are used in classification mode in a first step to identify the precipitating areas (precipitation detection) and in regression mode in a second step to estimate the rainfall intensity at the ground (rain rate). The input predictors are geostationary satellite infrared (IR) brightness temperatures and Satellite Application Facility (SAF) nowcasting products which consist of cloud properties, such as cloud top height and cloud type. Additionally, a set of auxiliary location-describing input variables is employed. The output predictand is the ground-based instantaneous rain rate provided by the European-scale radar composite OPERA, that was additionally quality-controlled. We compare our results to a precipitation product which uses a single infrared (IR) channel for the rainfall retrieval. Specifically, we choose the operational PR-OBS-3 hydrology SAF product as a representative example for this type of approach.With generalized linear models, we show that we are able to substantially improve in terms of hits by considering more IR channels and cloud property predictors. Furthermore, we demonstrate the added value of using artificial neural networks to further improve prediction skill by additionally reducing false alarms. In the rain rate estimation, the indirect relationship between surface rain rates and the cloud properties measurable with geostationary satellites limit the skill of all models, which leads to smooth predictions close to the mean rainfall intensity. Probability matching is explored as a tool to recover higher order statistics to obtain a more realistic rain rate distribution.

Deep Learning for Precipitation Estimation from Satellite and Rain Gauges Measurements

“…Compared to previous ML studies [4,9,10], we have introduced a multiscale, multimodal and multi-task DL model for precipitation area detection and instantaneous rain rate estimation from geostationary satellite imagery and rain gauges.…”

“…The performance of our model is not only coming from its multimodality, but is also due to our careful choice of DL architecture. Indeed, where other studies used a shallow fully connected NN [4,10], we used a deep multiscale convolutional NN able to learn spatial dependence in its input at different scales.…”

“…We evaluated the ability of our model to detect precipitation and to accurately estimate their rate by computing various scores already used in previous studies [9,10]. The considered classification scores include the Probability Of Detection (POD), the False Alarm Ratio (FAR), the Probability of False Detection (POFD), the Accuracy (ACC), the Critical Success Index (CSI), the Gilbert Skill Score (GSS), the Heidke Skill Score (HSS), the Hanssen-Kuipers Discriminant (HKD) and the F1 score (F1).…”

“…For the regression scores (Table 6), we used the Mean Error (ME), the Mean Absolute Error (MAE), the Root Mean Squared Error (RMSE), the Reduction of Variance (RV), the Pearson Correlation Coefficient (PCORR) and the Spearman Rank Correlation (SCORR). Descriptive information about these scores is available in [10]. In the results section, all these scores were computed individually for each test gauge and then all of them were averaged.…”

“…Thanks to the near-infrared measurements from MSG, more recent studies have been developed to use exclusively this GEO satellite and benefit from his high temporal resolution (15 min for the complete disk scan and 5 min for the rapid scan of Europe) and spatial resolution (3 km at nadir). Using the MSG data, [9] developed a 24 h precipitation estimation method relying on random forest and [10] used a fully connected neural network.…”