Performance assessment of artificial neural networks and support vector regression models for stream flow predictions.

Water resources planning, development, and management need reliable forecasts of river flows. In past few decades, an important dimension has been introduced in the prediction of the hydrologic phenomenon through artificial intelligence-based modeling. In this paper, the performance of three artificial neural network (ANN) and four support vector regression (SVR) models was investigated to predict streamflows in the Upper Indus River. Results from ANN models using three different optimization techniques, namely Broyden-Fletcher-Goldfarb-Shannon, Conjugate Gradient, and Back Propagation algorithms, were compared with one another. A further comparison was made between these ANNs and four types of SVR models which were based on linear, polynomial, radial basis function, and sigmoid kernels. Past 30 years' monthly data for precipitation, temperature, and streamflow obtained from Pakistan Surface Water Hydrology Department Lahore were used for this purpose. Three types of input combinations with respect to the main input variables (temperature, precipitation, and stream flow) and several types of input combinations with respect to time lag were tested. The best input for ANN and SVR models was identified using correlation coefficient analysis and genetic algorithm. The performance of the ANN and SVR models was evaluated by mean bias error, Nash-Sutcliffe efficiency, root mean square error, and correlation coefficient. The efficiency of the Broyden-Fletcher-Goldfarb-Shannon-ANN model was found to be much better than that of other models, while the SVR model based on radial basis function kernel predicted stream flows with comparatively higher accuracy than the other kernels. Finally, long-term predictions of streamflow have been made by the best ANN model. It was found that stream flow of Upper Indus River has a decreasing trend.

Performance evaluation of 1-D numerical model HEC-RAS towards modeling sediment depositions and sediment flushing operations for the reservoirs.

Sediment inflow to the reservoir is a key factor to calculate life of the reservoir. Flushing is a very useful technique in order to enhance the life of reservoir. From the literature review, 14 reservoirs were considered where flushing has already been practiced and only 6 reservoirs were found successful in flushing. The others were found partially flushed. In this research, data of three successfully flushed reservoirs namely Baira reservoir (India), Gebidem reservoir (Switzerland), and Gmund reservoir (Austria) were used to run 1-D numerical model HEC-RAS. In the first phase, the longitudinal profiles of delta were modeled and calibrated with observed sediment depositions. In the presence of available data of discharges and respective reservoir levels, the modeled deltas were used for flushing the deltas. With the help of modeled delta and corresponding flushing discharges, sediment deposition and flushing durations were computed. The simulated flushed durations were 31, 102, and 180 h for Baira, Gebidem, and Gmund reservoirs, respectively. The simulated flushed durations were found close to observed durations. Hence, the use of 1-D numerical model HEC-RAS is encouraged for modeling of sediment deposits and flushing operations.

Possible options to slow down the advancement rate of Tarbela delta.

The pivot point of delta in Tarbela dam has reached at about 10.6 km from the dam face which may result in blocking of tunnels. Tarbela delta was modeled from 1979 to 2060 using hec-6 model. Initially, the model was calibrated for year 1999 and validated for years 2000, 2001, 2002, and 2006 by involving the data of sediment concentration, reservoir cross sections (73 range lines), elevation-area capacity curves, and inflows and outflows from the reservoir. Then, the model was used to generate future scenarios, i.e., run-1, run-2, and run-3 with pool levels; 428, 442, and 457 m, respectively, till 2060. Results of run-1 and run-2 showed advancement to choke the tunnels by 2010 and 2030, respectively. Finally, in run-3, the advancement was further delayed showing that tunnels 1 and 2 will be choked by year 2050 and pivot point will reach at 6.4 km from the dam face.

Investigation of temporal change in glacial extent of Chitral watershed using Landsat data.

Glaciers are also known as solid reservoirs, and in this regard, Pakistan is a blessed country to have enriched glaciers. The change in glacial extent becomes very crucial for rivers whose discharges are associated with glacier melt. Even a little change in the glacial extent may bring a significant change in the resulting river flows. Considering climate change scenarios, many researchers have predicted future flows in such catchments. But in almost all such studies, the reduction in the glaciers is not normally based on any rational. Therefore, research is needed in order to estimate how glaciers are actually behaving under the change of temperature and precipitations to better estimate the future flows. For this purpose, Chitral watershed was considered as the study area. The seasonal change in the snow extent was estimated by using MODIS data for various years that helped to identify the month with minimum glacial extent. With the help of remote sensing, unsupervised classification was performed to estimate the glacier area in Chitral watershed. The results show a definite receding trend with respect to time in the glaciers of the region for the past decade.

Performance assessment of aquatic macrophytes for treatment of municipal wastewater.

The objective of the study was to evaluate the performance of three different aquatic macrophytes for treatment of municipal wastewater collected from Taxila (Pakistan). A physical model of treatment plant was constructed and was operated for six experimental runs with each species of macrophyte. Every experimental run consist of thirty days period. Regular monitoring of influent and effluent concentrations were made during each experimental run. For the treatment locally available macrophyte species i.e. water hyacinth, duckweed & water lettuce were selected to use. To evaluate the treatment performance of each macrophyte, BOD5, COD, and Nutrients (Nitrogen and Phosphorus) were monitored in effluent from model at different detention time of every experimental run after ensuring steady state conditions. The average reduction of effluent value of each parameter using water hyacinth were 50.61% for BOD5, 46.38% for COD, 40.34% for Nitrogen and 18.76% for Phosphorus. For duckweed the average removal efficiency for selected parameters were 33.43% for BOD5, 26.37% for COD, 17.59% for Nitrogen and 15.25% for Phosphorus and for Water Lettuce the average removal efficiency were 33.43% for BOD5, 26.37% for COD, 17.59% for Nitrogen and 15.25% for Phosphorus. The mechanisms of pollutant removal in this system include both aerobic and anaerobic microbiological conversions, sorption, sedimentation, volatilization and chemical transformations. The rapid growth of the biomass was measured within first ten days detention time. It was also observed that performance of macrophytes is influenced by variation of pH and Temperature. A pH of 6-9 and Temperature of 15-38°C is most favorable for treatment of wastewater by macrophytes. The option of macrophytes for treatment of Municipal sewage under local environmental conditions can be explored by further verifying the removal efficiency under variation of different environmental conditions. Also this is need of time that macrophyte system should be used for treatment of wastewater because their performance is comparable to conventional wastewater treatment plants and also the system has very low O&M costs.

Environmental and socio-economic impacts of pipe drainage in Pakistan.

Many drainage schemes and salinity control projects have been executed world wide. Pipe drainage has widely been used in Pakistan, Egypt and India to control waterlogging. The impact of pipe drainage on land and water was evaluated in this paper using data of three pipe drainage projects in Pakistan namely Khushab Salinity Control and Reclamation Project, Fourth Drainage Project in Faisalabad and Swabi Salinity Control and Reclamation Project. Data by regular monitoring of these projects were collected. The effect of pipe drainage on water table depth at these three locations has been compared. Water quality and soil salinity improvement due to the pipe drainage has also been investigated. Data, related to water table depths and discharges from drain pipes/wells, was collected. Observation wells, installed at various places by the Water and Power Development Authority, were used for collection of this data. To evaluate the impact of the projects on salinity, soil samples from all the three locations were tested. A questionnaire was prepared to get the view of the people about the projects. It was revealed that in these areas, due to subsurface pipe drainage, the percentage of the abandoned land has been considerably decreased. Over drainage was observed in a few places of the projects. The farmers at such places were asked to change their cropping patterns. Ultimately, there has been an increase in area under cultivation, crop yields and cropping intensity in the projects' area.

Development of geomorphologic instantaneous unit hydrograph for a large watershed.

Hill torrents cause a lot of environmental and property damage in Pakistan every year. Proper assessment of direct runoff in the form of hill torrents is essential for protection of environment, property, and human life. In this paper, direct surface runoff hydrograph (DSRH) was derived for a large catchment using the geomorphologic instantaneous unit hydrograph concept. The catchment with hill torrent flows in semi-arid region of Pakistan was selected for this study. It was divided into series of linear cascades and hydrologic parameters required for Nash's conceptual model, and were estimated using geomorphology of the basin. Geomorphologic parameters were derived from satellite images of the basin and ERDAS and ArcGIS were used for data processing. Computer program was developed to systematically estimate the dynamic velocity, its related parameters by optimization and thereby to simulate the DSRH. The data regarding rainfall-runoff and satellite images were collected from Punjab Irrigation and Power Department, Pakistan. Model calibration and validation was made for 15 rainfall-runoff events. Ten events were used for calibration and five for validation. Model efficiency was found to be more than 90% and root mean square error to be about 5%. Impact of variation in model parameters (shape parameter and storage coefficient) on DSRH was investigated. For shape parameter, the number of linear cascades varied from 1 to 3 and it was found that the shaper parameter value of 3 produced the best DSRH. Various values of storage coefficient were used and it was observed that the value determined from geomorphology and the dynamic velocity produced the best results.