Search

Drought, a climate-related disaster impacting a variety of sectors, poses challenges for millions of people in South Asia. Accurate and complete drought information with a proper monitoring system is very important in revealing the complex nature of drought and its associated factors. In this regard, deep learning is a very promising approach for delineating the non-linear characteristics of drought factors. Therefore, this study aims to monitor drought by employing a deep learning approach with remote sensing data over South Asia from 2001–2016. We considered the precipitation, vegetation, and soil factors for the deep forwarded neural network (DFNN) as model input parameters. The study evaluated agricultural drought using the soil moisture deficit index (SMDI) as a response variable during three crop phenology stages. For a better comparison of deep learning model performance, we adopted two machine learning models, distributed random forest (DRF) and gradient boosting machine (GBM). Results show that the DFNN model outperformed the other two models for SMDI prediction. Furthermore, the results indicated that DFNN captured the drought pattern with high spatial variability across three penology stages. Additionally, the DFNN model showed good stability with its cross-validated data in the training phase, and the estimated SMDI had high correlation coefficient R2 ranges from 0.57~0.90, 0.52~0.94, and 0.49~0.82 during the start of the season (SOS), length of the season (LOS), and end of the season (EOS) respectively. The comparison between inter-annual variability of estimated SMDI and in-situ SPEI (standardized precipitation evapotranspiration index) showed that the estimated SMDI was almost similar to in-situ SPEI. The DFNN model provides comprehensive drought information by producing a consistent spatial distribution of SMDI which establishes the applicability of the DFNN model for drought monitoring.

Heat-health risk is a growing concern in many regions of China due to the more frequent occurrence of extremely hot weather. Spatial indexes based on various heat assessment frameworks can be used for the assessment of heat risks. In this study, we adopted two approaches&mdash, Crichton&rsquo, s risk triangle and heat vulnerability index (HVI) to identify heat-health risks in the Northern Jiangxi Province of China, by using remote sensing and socio-economic data. The Geographical Information System (GIS) overlay and principal component analysis (PCA) were separately used in two frameworks to integrate parameters. The results show that the most densely populated community in the suburbs, instead of city centers, are exposed to the highest heat risk. A comparison of two heat assessment mapping indicates that the distribution of HVI highlights the vulnerability differences between census tracts. In contrast, the heat risk index of Crichton&rsquo, s risk triangle has a prominent representation for regions with high risks. The stepwise multiple linear regression zero-order correlation coefficient between HVI and outdoor workers is 0.715, highlighting the vulnerability of this particular group. Spearman&rsquo, s rho nonparametric correlation and the mean test reveals that heat risk index is strongly correlated with HVI in most of the main urban regions in the study area, with a significantly lower value than the latter. The analysis of variance shows that the distribution of HVI exhibits greater variety across urban regions than that of heat risk index. Our research provides new insight into heat risk assessment for further study of heat health risk in developing countries.

Drought has serious consequences on terrestrial ecosystems, particularly for their carbon and water processes. As an important indicator to examine the balance of ecosystem water and carbon cycles, ecosystem water use efficiency (WUE) has been widely used to investigate ecosystem responses to drought. However, the response of WUE to drought and the role of different ecosystem processes in controlling the response of WUE to drought are not well studied. In this paper, we used four WUE datasets from different remote sensing-driven (RS-driven) models and three drought indices (Standardized Precipitation Evapotranspiration Index, soil moisture anomaly index and water storage anomaly-based drought index) to comprehensively investigate the response of WUE to drought and its dominant ecosystem processes during the period of 2001-2018. The results showed the WUE datasets from four different RS-driven models had discrepancies in WUE temporal trends, particularly in tropical and subtropical forest and semi-arid regions. The Spearman correlation analysis revealed that the positive correlations between WUE and drought accounted for more than half of global vegetated lands, while negative relationship mainly occurred in the high latitude regions. We further explored the dominant ecosystem processes (represented by GPP and ET) in controlling WUE response to drought, and found ET controlled WUE-drought relationship in the high latitude areas and semi-arid/sub-humid regions, while GPP dominated it in tropical forest regions. Additionally, the effects of GPP and ET on controlling WUE response to drought were examined to change with different drought indices, especially in the semi-arid regions. Our study suggests multi-model analysis tend to reduce uncertainties in analyzing WUE response to drought caused by a single WUE data. Moreover, our results highlight the different role of ecosystem processes in controlling WUE response to drought and provide new information for the underlying mechanism of drought impacts on ecosystem water and carbon cycles.