Search

Hydrological modelling in karst environments is a difficult task due to the inherent complexity of the karst system and the usual lack of information about its geometrical description. The hydrological processes of karst systems in Mediterranean semiarid environments are particularly difficult to simulate mathematically due to the pattern of long dry periods and short wet periods. In this study, we tested the ability of the open-source SWAT hydrologic modelling code to simulate the behaviour and hydrologic output of a karstic watershed in a Mediterranean semiarid environment (SE Spain). Calibration and first validation were accomplished using a 20-year and 10-year record of stream water discharge, respectively, at the catchment outlet. Additional testing of the model was accomplished using groundwater discharge data from four natural springs in the watershed and by comparing the results of the SWAT model with the SIMPA model (a hydrological model used by the Spanish national water authority to simulate the water balance among others). Likewise, an adapted form of the Nash-Sutcliffe Efficiency (NSE) index for arid environments, ANSE, is presented. Based on results, the simulated behaviour was good and very good (with ANSE values of 0.96 and 0.78 for the calibration and validation periods respectively). SWAT and SIMPA results provide spatial distributions of the main hydrological processes of the watershed, such as aquifer recharge, actual evapotranspiration, and surface runoff, being verified as useful tools for water policy managers in karstic environments.Keywords:Water balance; karst aquifer springs; SWAT; SIMPA; ANSE index; Mediterranean karstic catchment.

Owing to the extensive global dependency on groundwater and associated increasing water demand, the global groundwater level is declining rapidly. In the case of Islamabad, Pakistan, the groundwater level has lowered five times over the past five years due to extensive pumping by various departments and residents to meet the local water requirements. To address this, water reservoirs and sources need to be delineated, and potential recharge zones are highlighted to assess the recharge potential. Therefore, the current study utilizes an integrated approach based on remote sensing (RS) and GIS using the influence factor (IF) technique to delineate potential groundwater recharge zones in Islamabad, Pakistan. Soil map of Pakistan, Landsat 8TM satellite data, digital elevation model (ASTER DEM), and local geological map were used in the study for the preparation of thematic maps of 15 key contributing factors considered in this study. To generate a combined groundwater recharge map, rate and weightage values were assigned to each factor representing their mutual influence and recharge capabilities. To analyze the final combined recharge map, five different assessment analogies were used in the study: poor, low, medium, high, and best. The final recharge potential map for Islamabad classifies 15% (136.8 km2) of the region as the “best” zone for extracting groundwater. Furthermore, high, medium, low, and poor ranks were assigned to 21%, 24%, 27%, and 13% of the region with respective areas of 191.52 km2, 218.88 km2, 246.24 km2, and 118.56 km2. Overall, this research outlines the best to least favorable zones in Islamabad regarding groundwater recharge potentials. This can help the authorities devise mitigation strategies and preserve the natural terrain in the regions with the best groundwater recharge potential. This is aligned with the aims of the interior ministry of Pakistan for constructing small reservoirs and ponds in the existing natural streams and installing recharging wells to maintain the groundwater level in cities. Other countries can expand upon and adapt this study to delineate local groundwater recharge potentials.

More than 8000 pedestrians were killed due to road crashes in Australia over the last 30 years. Pedestrians are assumed to be the most vulnerable users of roads. This susceptibility of pedestrians to road crashes conflicts with sustainable transportation objectives. It is critical to know the causes of pedestrian injuries in order to enhance the safety of these vulnerable road users. To achieve this, traditional statistical models are used frequently. However, they have been criticized for their inflexibility in handling outliers and missing or noisy data, and their strict pre-assumptions. This study applied an advanced machine learning algorithm, a Bayesian neural network, which has the characters of both Bayesian theory and neural networks. Several structures of this model were built, and the best structure was selected, which included three hidden neuron layers—sixteen hidden nodes in the first layer and eight hidden nodes in the second and third layers. The performance of this model was compared with the performances of some other machine learning techniques, including standard Bayesian networks, a standard neural network, and a random forest model. The Bayesian neural network model outperformed the other models. In addition, a study on the importance of the features showed that the individuals’ characteristics, time, and circumstantial factors were essential. They greatly increased model performance if the model used them. This research lays the groundwork for using machine learning approaches to alleviate pedestrian deaths caused by road accidents.