As the successor of the Tropical Rainfall Measuring Mission (TRMM), the Global Precipitation Measurement (GPM) mission significantly improves the spatial resolution of precipitation estimates from 0.25° to 0.1°. The present study analyzed the error structures of Integrated Multisatellite Retrievals for GPM (IMERG) monthly precipitation products over Mainland China from March 2014 to February 2015 using gauge measurements at multiple spatiotemporal scales. Moreover, IMERG products were also compared with TRMM 3B43 products. The results show that: (1) overall, IMERG can capture the spatial patterns of precipitation over China well. It performs a little better than TRMM 3B43 at seasonal and monthly scales; (2) the performance of IMERG varies greatly spatially and temporally. IMERG performs better at low latitudes than at middle latitudes, and shows worse performance in winter than at other times; (3) compared with TRMM 3B43, IMERG significantly improves the estimation accuracy of precipitation over the Xinjiang region and the Qinghai-Tibetan Plateau, especially over the former where IMERG increases Pearson correlation coefficient by 0.18 and decreases root-mean-square error by 54.47 mm for annual precipitation estimates. However, most IMERG products over these areas are unreliable; and (4) IMERG shows poor performance in winter as TRMM 3B43 even if GPM improved its ability to sense frozen precipitation. Most of them over North China are unreliable during this period.

Labeled samples are important in achieving land cover change detection (LCCD) tasks via deep learning techniques with remote sensing images. However, labeling samples for change detection with bitemporal remote sensing images is labor-intensive and time-consuming. Moreover, manually labeling samples between bitemporal images requires professional knowledge for practitioners. To address this problem in this article, an iterative training sample augmentation (ITSA) strategy to couple with a deep learning neural network for improving LCCD performance is proposed here. In the proposed ITSA, we start by measuring the similarity between an initial sample and its four-quarter-overlapped neighboring blocks. If the similarity satisfies a predefined constraint, then a neighboring block will be selected as the potential sample. Next, a neural network is trained with renewed samples and used to predict an intermediate result. Finally, these operations are fused into an iterative algorithm to achieve the training and prediction of a neural network. The performance of the proposed ITSA strategy is verified with some widely used change detection deep learning networks using seven pairs of real remote sensing images. The excellent visual performance and quantitative comparisons from the experiments clearly indicate that detection accuracies of LCCD can be effectively improved when a deep learning network is coupled with the proposed ITSA. For example, compared with some state-of-the-art methods, the quantitative improvement is 0.38%–7.53% in terms of overall accuracy. Moreover, the improvement is robust, generic to both homogeneous and heterogeneous images, and universally adaptive to various neural networks of LCCD. The code will be available at https://github.com/ImgSciGroup/ITSA .

Spatially downscaling satellite precipitation products have been performed on annual and monthly precipitation. Accurate downscaling on daily precipitation remains a challenge due to the limitation of the downscaling assumption, the large spatial discontinuity of daily precipitation, and the relatively poor quality of satellite‐derived daily precipitation product. In this study, an integrated downscaling‐fusion framework was proposed and used to downscale satellite‐derived daily precipitation. First, a spatio‐temporal downscaling scheme is applied to produce preliminary downscaled daily precipitation. The accuracy of the derived preliminary results is then boosted by merging with daily gauge observations using an ensemble fusion method. The performance of the proposed framework was tested and evaluated by downscaling the Integrated Multi‐satellite Retrievals for Global Precipitation Measurement (IMERG) daily precipitation data from 0.1 to 0.01° over eastern and central China for the period of 2015–2016. The results showed that (a) the downscaling scheme accurately mapped the spatio‐temporal variation in daily precipitation, and the preliminary downscaled results perfectly maintained the accuracy of the original IMERG data; (b) the fused results were much more accurate than the original IMERG data, decreasing the root‐mean‐square errors (RMSEs) by 22, 10, and 18% at daily, monthly, and annual timescales, respectively, for the whole period; and (c) the fused daily precipitation data considerably strengthened the detection of rain/no rain area compared with the original IMERG daily precipitation data, with a 17% reduction in the inconsistency index.

Many research papers have reported problems with existing fusion techniques. The most significant problem is that fusion images produce spectral distortion when they retain high spatial resolution. Independent component analysis (ICA) can eliminate low-order and also high-order redundancy for data. According to statistical theory, the most important information from data is always included in the statistical characteristic of high order. Therefore, independent components (ICs) resulting from the transform of the ICA provide information on the data that is otherwise hidden in the large data set, and describe the essential structure of the data. For a colour or false colour composite image, the ICs represent the main body, spectral and spatial detail information. In this study, a novel remote sensing image fusion method based on ICA is presented. By replacing the main body IC of a multispectral image with the vector of a panchromatic (PAN) image, the new ICs contain both spatial information of the PAN image and spectral characteristics of the multispectral image. Then an inverse ICA transformation (IIT) is performed to attain the fusion image. IKONOS and Enhanced Thematic Mapper Plus (ETM + ) images are used to evaluate the proposed method and others (hue–saturation–value (HSV), principal component analysis and wavelets). The fusion results are compared graphically, visually and statistically, and show that the proposed method can retain spatial and spectral information simultaneously, and has a better balance between them.

The lack of official statistics makes it difficult to assess Venezuela's economic situation during the socioeconomic crisis. In this study, we used Visible Infrared Imaging Radiometer Suite (VIIRS) nighttime light images to evaluate Venezuela's crisis. The Hodrick—Prescott filter was used to decompose the sum of urban light (SUL) into two components, the sum of urban light trend (SULT) and the sum of urban light cycle. Then we proposed an index of nighttime light change ratio (NLCR) to estimate Venezuelan nighttime light dynamics. We found that Venezuela has lost 30.37% of its SULT from April 2012 to December 2018. The regression analysis shows that Venezuelan SULT had a strong relationship to a number of socioeconomic indicators: the SULT is positively correlated to crude oil production with R $^2$ of 0.9159, and negatively correlated to dollar exchange rate with R $^2$ of 0.9516, and negatively correlated to the number of asylum seekers with R2 of 0.8384. We also found that among the three states with largest nighttime light decrease, economy of two states was dominated by agriculture and that of one state was dominated by the oil industry. In the pixel analysis, compared with the urban cores, the suburbs of urban cores of 12 main cities have a higher percentage of SULT-increased areas. Around the Venezuela-Colombia border, the SULT decreased in the Venezuelan side but increased in the Colombian side. Our analysis suggests that nighttime light imagery can help to assess Venezuela's situation during the crisis.