We propose a novel spatiotemporal fusion method based on deep convolutional neural networks (CNNs) under the application background of massive remote sensing data. In the training stage, we build two five-layer CNNs to deal with the problems of complicated correspondence and large spatial resolution gaps between MODIS and Landsat images. Specifically, we first learn a nonlinear mapping CNN between MODIS and low-spatial-resolution (LSR) Landsat images and then learn a super-resolution CNN between LSR Landsat and original Landsat images. In the prediction stage, instead of directly taking the outputs of CNNs as the fusion result, we design a fusion model consisting of high-pass modulation and a weighting strategy to make full use of the information in prior images. Specifically, we first map the input MODIS images to transitional images via the learned nonlinear mapping CNN and further improve the transitional images to LSR Landsat images via the fusion model; then, via the learned SR CNN, the LSR Landsat images are supersolved to transitional images, which are further improved to Landsat images via the fusion model. Compared with the previous learning-based fusion methods, mainly referring to the sparse-representation-based methods, our CNNs-based spatiotemporal method has the following advantages: 1) automatically extracting effective image features; 2) learning an end-to-end mapping between MODIS and LSR Landsat images; and 3) generating more favorable fusion results. To examine the performance of the proposed fusion method, we conduct experiments on two representative Landsat-MODIS datasets by comparing with the sparse-representation-based spatiotemporal fusion model. The quantitative evaluations on all possible prediction dates and the comparison of fusion results on one key date in both visual effect and quantitative evaluations demonstrate that the proposed method can generate more accurate fusion results.

Land evaporation (ET) plays a crucial role in the hydrological and energy cycle. However, the widely used model-based products, even though helpful, are still subject to great uncertainties due to imperfect model parameterizations and forcing data. The lack of available observed data has further complicated estimation. Hence, there is an urgency to define the global proxy land ET with lower uncertainties for climate-induced hydrology and energy change. This study has combined three existing model-based products – the fifth-generation ECMWF reanalysis (ERA5), Global Land Data Assimilation System Version 2 (GLDAS2), and the second Modern-Era Retrospective analysis for Research and Applications (MERRA-2) – to obtain a single framework of a long-term (1980–2017) daily ET product at a spatial resolution of 0.25^∘. Here, we use the reliability ensemble averaging (REA) method, which minimizes errors using reference data, to combine the three products over regions with high consistencies between the products using the coefficient of variation (CV). The Global Land Evaporation Amsterdam Model Version 3.2a (GLEAM3.2a) and flux tower observation data were selected as the data for reference and evaluation, respectively. The results showed that the merged product performed well over a range of vegetation cover scenarios. The merged product also captured the trend of land evaporation over different areas well, showing the significant decreasing trend in the Amazon Plain in South America and Congo Basin in central Africa and the increasing trend in the east of North America, west of Europe, south of Asia and north of Oceania. In addition to demonstrating a good performance, the REA method also successfully converged the models based on the reliability of the inputs. The resulting REA data can be accessed at https://doi.org/10.5281/zenodo.4595941 (Lu et al., 2021).

Assessing the long-term precipitation changes is of utmost importance for understanding the impact of climate change. This study investigated the variability of extreme precipitation events over Pakistan on the basis of daily precipitation data from 51 weather stations from 1980-2016. The non-parametric Mann–Kendall, Sen’s slope estimator, least squares method, and two-tailed simple t-test methods were used to assess the trend in eight precipitation extreme indices. These indices were wet days (R1 ≥1 mm), heavy precipitation days (R10 ≥ 10 mm), very heavy precipitation days (R20 ≥ 20 mm), severe precipitation (R50 ≥ 50 mm), very wet days (R95p) defining daily precipitation ≥ 95 percentile, extremely wet days (R99p) defining daily precipitation ≥ 99 percentile, annual total precipitation in wet days (PRCPTOT), and mean precipitation amount on wet days as simple daily intensity index (SDII). The study is unique in terms of using high stations’ density, extended temporal coverage, advanced statistical techniques, and additional extreme indices. Furthermore, this study is the first of its kind to detect abrupt changes in the temporal trend of precipitation extremes over Pakistan. The results showed that the spatial distribution of trends in different precipitation extreme indices over the study region increased as a whole; however, the monsoon and westerlies humid regions experienced a decreasing trend of extreme precipitation indices during the study period. The results of the sequential Mann–Kendall (SqMK) test showed that all precipitation extremes exhibited abrupt dynamic changes in temporal trend during the study period; however, the most frequent mutation points with increasing tendency were observed during 2011 and onward. The results further illustrated that the linear trend of all extreme indices showed an increasing tendency from 1980- 2016. Similarly, for elevation, most of the precipitation extremes showed an inverse relationship, suggesting a decrease of precipitation along the latitudinal extent of the country. The spatiotemporal variations in precipitation extremes give a possible indication of the ongoing phenomena of climate change and variability that modified the precipitation regime of Pakistan. On the basis of the current findings, the study recommends that future studies focus on underlying physical and natural drivers of precipitation variability over the study region.

Various state-of-the-art gridded satellite precipitation products (GPPs) have been derived from remote sensing and reanalysis data and are widely used in hydrological studies. An assessment of these GPPs against in-situ observations is necessary to determine their respective strengths and uncertainties. GPPs developed from satellite observations as a primary source were compared to in-situ observations, namely the Climate Hazard group Infrared Precipitation with Stations (CHIRPS), Multi-Source Weighted-Ensemble Precipitation (MSWEP), Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks-Climate Data Record (PERSIANN-CDR) and Tropical Rainfall Measuring Mission (TRMM) multi-satellite precipitation analysis (TMPA). These products were compared to in-situ data from 51 stations, spanning 1998–2016, across Pakistan on daily, monthly, annual and interannual time scales. Spatiotemporal climatology was well captured by all products, with more precipitation in the north eastern parts during the monsoon months and vice-versa. Daily precipitation with amount larger than 10 mm showed significant (95%, Kolmogorov-Smirnov test) agreement with the in-situ data, especially TMPA, followed by CHIRPS and MSWEP. At monthly scales, there were significant correlations (R) between the GPPs and in-situ records, suggesting similar dynamics; however, statistical metrics suggested that the performance of these products varies from north towards south. Temporal agreement on an interannual scale was higher in the central and southern parts which followed precipitation seasonality. TMPA performed the best, followed in order by CHIRPS, MSWEP and PERSIANN-CDR.

Actual evapotranspiration (ET) is an essential variable in the hydrological process, linking carbon, water, and energy cycles. Global ET has significantly changed in the warming climate. Although the increasing vapor pressure deficit (VPD) enhances atmospheric water demand due to global warming, it remains unclear how the dynamics of ET are affected. In this study, using multiple datasets, we disentangled the relative contributions of precipitation, net radiation, air temperature (T₁), VPD, and wind speed on the annual ET linear trend using an advanced separation method that considers the Budyko framework. We found that the precipitation variability dominantly controls global ET in the dry climates, while the net radiation has substantial control over ET in the tropical regions, and VPD impacts ET trends in the boreal mid-latitude climate. The critical role of VPD in controlling ET trends is particularly emphasized due to its influence in controlling the carbon–water–energy cycle.