Cyanobacterial harmful algal blooms (CyanoHABs) threaten ecosystem functioning and human health at both regional and global levels, and this threat is likely to become more frequent and severe under climate change. Predictive information can help local water managers to alleviate or manage the adverse effects posed by CyanoHABs. Previous works have led to various approaches for predicting cyanobacteria abundance by feeding various environmental variables into statistical models or neural networks. However, these models alone may have limited predictive performance owing to their inability to capture extreme situations. In this paper, we consider the possibility of a hybrid approach that leverages the merits of these methods by integrating a statistical model with a deep-learning model. In particular, the autoregressive integrated moving average (ARIMA) and long short-term memory (LSTM) were used in tandem to better capture temporal patterns of highly dynamic observations. Results show that the proposed ARIMA-LSTM model exhibited the promising potential to outperform the state-of-the-art baseline models for CyanoHAB prediction in highly variable time-series observations, characterized by nonstationarity and imbalance. The predictive error of the mean absolute error and root mean square error, compared with the best baseline model, were largely reduced by 12.4% and 15.5%, respectively. This study demonstrates the potential for the hybrid model to assist in cyanobacterial risk assessment and management, especially in shallow and eutrophic waters.

We assessed the present and future climatologies of mean summer monsoon over South Asia using a high resolution regional climate model (RegCM4) with a 25 km horizontal resolution. In order to evaluate the performance of the RegCM4 for the reference period (1976–2005) and for the far future (2070–2099), climate change projections under two greenhouse gas representative concentration pathways (RCP4.5 and RCP8.5) were made, the lateral boundary conditions being provided by the geophysical fluid dynamic laboratory global model (GFDL-ESM2M). The regional climate model (RCM) improves the simulation of seasonal mean temperature and precipitation patterns compared to driving global climate model (GCM) during present-day climate conditions. The regional characteristic features of South Asian summer monsoon (SASM), like the low level jet stream and westerly flow over the northern the Arabian Sea, are well captured by the RegCM4. In spite of some discrepancies, the RegCM4 could simulate the Tibetan anticyclone and the direction of the tropical easterly jet reasonably well at 200 hPa. The projected temperature changes in 2070–2099 relative to 1976–2005 for GFDL-ESM2M show increased warming compared to RegCM4. The projected patterns at the end of 21st century shows an increase in precipitation over the Indian Peninsula and the Western Ghats. The possibilities of excessive precipitation include increased southwesterly flow in the wet period and the effect of model bias on climate change. However, the spatial patterns of precipitation are decreased in intensity and magnitude as the monsoon approaches the foothills of the Himalayas. The RegCM4-projected dry conditions over northeastern India are possibly related to the anomalous anticyclonic circulations in both scenarios.

China launched the sponge city (SPC) initiative in 2013 to reduce municipal stormwater runoff. The design criteria are mainly the annual comprehensive runoff coefficient (ACRC) regulated in a design guideline. Numerous SPC alternatives with varied low-impact development (LID) measures can be designed to meet the ACRC. Obviously, the optimization of SPC design is significant. This study provides an approach to SPC design optimization that applies an optimized module of SUSTAIN to simulate SPC performance over a 10-year period. The targeted volume reduction was derived from the SWMM model and corresponded to the ACRC criteria. Based on the reduction, the minimal cost and cost-effectiveness analyses were conducted. The proposed approach was applied to the Lincang Technology Innovation Park (LCTIP) as a test case. Three scenarios were analyzed: The original design implemented on the site, the landscape improved design, and the most economical design. The results indicated that the optimized alternative may save up to 12.3% of the cost while meeting that ACRC value. The approach improves upon SPC design, particularly with regards to flood control. The present research will help decision makers to develop and select the most appropriate SPC design that is most cost-effective.