Groundwater Drought Analysis under Data Scarcity: The Case of the Salento Aquifer (Italy)

Salento is a regional coastal karst aquifer located in Southern Italy with a highly complex geological, geomorphological, and hydrogeological structure. High and unruly exploitation of groundwater from licensed and unlicensed wells for irrigation and drinking purposes affects groundwater, with consequent degradation of its qualitative and quantitative status. The increased frequency of meteorological droughts and rising temperatures may only worsen the already compromised situation. The absence of complete and enduring monitoring of groundwater levels prevents the application of some methodologies, which require long time series. The analysis of climate indexes to describe the groundwater level variation is a possible approach under data scarcity. However, this approach may not be obvious for complex aquifers (in terms of scale, intrinsic properties, and boundary conditions) where the response of the groundwater to precipitation is not necessarily linear. Thus, the proposed research deals with the assessment of the response of the Salento aquifer to precipitation variability based on correlations between the Standardized Precipitation Index (SPI) and Standardized Precipitation and Evapotranspiration Index (SPEI) and groundwater levels for nine monitoring wells from July 2007 to December 2011. The study aims at evaluating the ability of the above indicators to explain the behavior of groundwater on complex aquifers. Moreover, it has the general aim to verify their more general reliable application. Results of three different correlation factors outline direct and statistically significant correlations between the time series. They describe the Salento aquifer as a slow filter, with a notable inertial behavior in response to meteorological events. The SPI 18-months demonstrates to be a viable candidate to predict the groundwater response to precipitation variability for the Salento aquifer.

CRNNTL: Convolutional Recurrent Neural Network and Transfer Learning for QSAR Modeling in Organic Drug and Material Discovery

Molecular latent representations, derived from autoencoders (AEs), have been widely used for drug or material discovery over the past couple of years. In particular, a variety of machine learning methods based on latent representations have shown excellent performance on quantitative structure–activity relationship (QSAR) modeling. However, the sequence feature of them has not been considered in most cases. In addition, data scarcity is still the main obstacle for deep learning strategies, especially for bioactivity datasets. In this study, we propose the convolutional recurrent neural network and transfer learning (CRNNTL) method inspired by the applications of polyphonic sound detection and electrocardiogram classification. Our model takes advantage of both convolutional and recurrent neural networks for feature extraction, as well as the data augmentation method. According to QSAR modeling on 27 datasets, CRNNTL can outperform or compete with state-of-art methods in both drug and material properties. In addition, the performances on one isomers-based dataset indicate that its excellent performance results from the improved ability in global feature extraction when the ability of the local one is maintained. Then, the transfer learning results show that CRNNTL can overcome data scarcity when choosing relative source datasets. Finally, the high versatility of our model is shown by using different latent representations as inputs from other types of AEs.

GSTools v1.3: A toolbox for geostatistical modelling in Python

Geostatistics as a subfield of statistics accounts for the spatial correlations encountered in many applications of e.g. Earth Sciences. Valuable information can be extracted from these correlations, also helping to address the often encountered burden of data scarcity. Despite the value of additional data, the use of geostatistics still falls short of its potential. This problem is often connected to the lack of user-friendly software hampering the use and application of geostatistics. We therefore present GSTools, a Python-based software suite for solving a wide range of geostatistical problems. We chose Python due to its unique balance between usability, flexibility, and efficiency and due to its adoption in the scientific community. GSTools provides methods for generating random fields, it can perform kriging and variogram estimation and much more. We demonstrate its abilities by virtue of a series of example application detailing their use.