Solving transparency in drought forecasting using attention models

An enhanced drought forecasting in coastal arid regions using deep learning approach with evaporation index

Coastal arid regions are similar to deserts, where it receives significantly less rainfall, less than 10 cm. Perhaps the world's worst natural disaster, coastal area droughts, can only be detected using reliable monitoring systems. Creating a reliable drought forecast model and figuring out how well various models can analyze drought factors in coastal arid regions are two of the biggest obstacles in this field. Different time-series methods and machine-learning models have traditionally been utilized in forecasting strategies. Deep learning is promising when describing the complex interplay between coastal drought and its contributing variables. Considering the possibility of enhancing our understanding of drought features, applying deep learning approaches has yet to be tried widely. The current investigation employs a deep learning strategy. Coastal Drought indices are commonly used to comprehend the situation better; hence the Standard Precipitation Evaporation Index (SPEI) was used since it incorporates temperatures and precipitation into its computation. An integrated coastal drought monitoring model was presented and validated using convolutional long short-term memory with self-attention (SA-CLSTM). The Climatic Research Unit (CRU) dataset, which spans 1901-2018, was mined for the drought index and predictor data. To learn how LSTM forecasting could enhance drought forecasting, we analyzed the findings regarding numerous drought parameters (drought severity, drought category, or geographic variation). The model's ability to predict drought intensity was assessed using the Coefficient of Determination (R2), the Root Mean Square Error (RMSE), and the Mean Absolute Error (MAE). Both the SPEI 1 and SPEI 3 examples had R2 values more than 0.99 for the model. The range of predicted outcomes for each drought group was analyzed using a multi-class Receiver Operating Characteristic based Area under Curves (ROC-AUC) method. The research showed that the AUC for SPEI 1 was 0.99 and for SPEI 3, 0.99. The study's results indicate progress over machine learning models for one month in advance, accounting for various drought conditions. This work's findings may be used to mitigate drought, and additional improvement can be achieved by testing other models.

Heterogeneous catalysis mediated by light, electricity and enzyme via machine learning: Paradigms, applications and prospects

Energy crisis and environmental pollution have become the bottleneck of human sustainable development. Therefore, there is an urgent need to develop new catalysts for energy production and environmental remediation. Due to the high cost caused by blind screening and limited valuable computing resources, the traditional experimental methods and theoretical calculations are difficult to meet with the requirements. In the past decades, computer science has made great progress, especially in the field of machine learning (ML). As a new research paradigm, ML greatly accelerates the theoretical calculation methods represented by first principal calculation and molecular dynamics, and establish the physical picture of heterogeneous catalytic processes for energy and environment. This review firstly summarized the general research paradigms of ML in the discovery of catalysts. Then, the latest progresses of ML in light-, electricity- and enzyme-mediated heterogeneous catalysis were reviewed from the perspective of catalytic performance, operating conditions and reaction mechanism. The general guidelines of ML for heterogeneous catalysis were proposed. Finally, the existing problems and future development trend of ML in heterogeneous catalysis mediated by light, electricity and enzyme were summarized. We highly expect that this review will facilitate the interaction between ML and heterogeneous catalysis, and illuminate the development prospect of heterogeneous catalysis.