There is a drastic geographic imbalance in available global streamflow gauge and catchment property data, with additional large variations in data characteristics. As a result, models calibrated in one region cannot normally be migrated to another without significant modifications. Currently in these regions, non‐transferable machine learning models are habitually trained over small local data sets. Here we show that transfer learning (TL), in the senses of weight initialization and weight freezing, allows long short‐term memory (LSTM) streamflow models that were pretrained over the conterminous United States (CONUS, the source data set) to be transferred to catchments on other continents (the target regions), without the need for extensive catchment attributes available at the target location. We demonstrate this possibility for regions where data are dense (664 basins in Great Britain), moderately dense (49 basins in central Chile), and scarce with only remotely sensed attributes available (5 basins in China). In both China and Chile, the TL models showed significantly elevated performance compared to locally trained models using all basins. The benefits of TL increased with the amount of available data in the source data set, and seemed to be more pronounced with greater physiographic diversity. The benefits from TL were greater than from pretraining LSTM using the outputs from an uncalibrated hydrologic model. These results suggest hydrologic data around the world have commonalities which could be leveraged by deep learning, and synergies can be had with a simple modification of the current workflows, greatly expanding the reach of existing big data. Finally, this work diversified existing global streamflow benchmarks.
This paper presents an analysis of trends in six drought variables at 566 stations across India over the period 1901-2002. Six drought variables were computed using standardized precipitation index (SPI). The Mann-Kendall (MK) trend test and Sen’s slope estimator were used for trend analysis of drought variables. Discrete wavelet transform (DWT) was used to identify the dominant periodic components in trends, whereas the significance of periodic components was examined using continuous wavelet transform (CWT) based global wavelet spectrum (GWS). Our results show an increasing trend in droughts in eastern, northeastern and extreme southern regions, and a decreasing trend in northern and southern regions of the country. The periodic component influencing the trend was 2-4 years in south, 4-8 years in west, east and northeast, 8-64 years in central parts and 32-128 years in north; however, most of the periodic components were not statistically significant.