Application of Artificial Neural Networks to Predict Total Dissolved Solids at the Karaj Dam

We applied multilayer perceptron (MLP) and radial basis function (RBF) neural networks using data from two water quality monitoring stations at the Karaj Dam in Iran. Input data were calcium ions (Ca²⁺), magnesium ions (Mg²⁺), sodium ions (Na⁺), chloride ions (Cl^?), sulfate (), and pH, and the output data were total dissolved solids (TDS). An MLP with one hidden layer containing eight neurons was selected for the upstream water quality station using normalized input data. We developed a second MLP neural network for the downstream station with one hidden layer containing 10 neurons in the hidden layer using normalized input data. Considering applying normalized input data and one hidden layer, the coefficient of determination (R ²) and index of agreement (IA) between the observed and the predicted data for the upstream and downstream monitoring stations using the MLP neural networks were 0.985, 0.84, 0.99, and 0.92, respectively. The RBF neural network with 100 neurons in its hidden layer reached the minimum errors between the observed and the predicted results in upstream and downstream stations. The R ² between observed and predicted data for upstream and downstream monitoring stations for the RBF was 0.999 and 0.998, respectively. Data normalization improved the performance of the MLP neural networks. Sensitivity analysis indicated that magnesium is the most effective water quality parameter for predicting TDS, and sulfate is the second most effective water quality parameter affecting TDS prediction at the Karaj Dam.     

Flood Forecasting Using Neural Computing Techniques and Conceptual Class Segregation

Various neural networks models are developed and applied for flood forecasting at Sangye station (no. 1) of the Bocheong Stream catchment, which is one of the International Hydrological Program's representative catchments, Republic of Korea. The neural networks models (NNMs) are multilayer perceptron‐neural networks model (MLP‐NNM), generalized regression neural networks model (GRNNM), and Kohonen self‐organizing feature maps neural networks model (KSOFM‐NNM). Data used for model training and testing are divided into two groups: such as floods and typhoon events. Single conventional application and class segregation implementation are applied to evaluate the neural networks models. KSOFM‐NNM forecasts flood discharge more accurately than do MLP‐NNM and GRNNM for the testing data of Methods I and II for single conventional application and class segregation implementation. This study shows that class segregation can capture the dynamics of different physical processes and overcome the difficulties using single conventional application of neural networks models.     

基于人工神经网络的绥化市县域经济发展水平分析

选取8个经济指标，运用人工神经网络（ANN）的理论和方法，构建应用最为广泛的BP网络模型，对2004年绥化市10个县市的经济发展水平进行了评价。结果表明，绥化市县域经济发展水平差异十分显著，其中肇东等3县域属于高水平类型，海伦等4个县域为中等类型，明水等3个县域属于落后类型。     

Modeling biological oxygen demand of the Melen River in Turkey using an artificial neural network technique

Artificial neural networks (ANNs) are being used increasingly to predict and forecast water resources' variables. The feed-forward neural network modeling technique is the most widely used ANN type in water resources applications. The main purpose of the study is to investigate the abilities of an artificial neural networks' (ANNs) model to improve the accuracy of the biological oxygen demand (BOD) estimation. Many of the water quality variables (chemical oxygen demand, temperature, dissolved oxygen, water flow, chlorophyll a and nutrients, ammonia, nitrite, nitrate) that affect biological oxygen demand concentrations were collected at 11 sampling sites in the Melen River Basin during 2001-2002. To develop an ANN model for estimating BOD, the available data set was partitioned into a training set and a test set according to station. In order to reach an optimum amount of hidden layer nodes, nodes 2, 3, 5, 10 were tested. Within this range, the ANN architecture having 8 inputs and 1 hidden layer with 3 nodes gives the best choice. Comparison of results reveals that the ANN model gives reasonable estimates for the BOD prediction.     

Field Evaluation of DRAINMOD 5.1 Under a Cold Climate: Simulation of Daily Midspan Water Table Depths and Drain Outflows1

Abstract: The hydrologic performance of DRAINMOD 5.1 was assessed for the southern Quebec region considering freezing/thawing conditions. A tile drained agricultural field in the Pike River watershed was instrumented to measure tile drainage volumes. The model was calibrated using water table depth and subsurface flow data over a two‐year period, while another two‐year dataset served to validate the model. DRAINMOD 5.1 accurately simulated the timing and magnitude of subsurface drainage events. The model also simulated the pattern of water table fluctuations with a good degree of accuracy. The R² between the observed and simulated daily WTD for calibration was >0.78, and that for validation was 0.93. The corresponding coefficients of efficiency (E) were >0.74 and 0.31. The R² and E values for calibration/validation of subsurface flow were 0.73/0.48 and 0.72/0.40, respectively. DRAINMOD simulated monthly subsurface flow quite accurately (E > 0.82 and R² > 0.84). The model precisely simulated daily/monthly drain flow over the entire year, including the winter months. Thus DRAINMOD 5.1 performed well in simulating the hydrology of a cold region.     

A Weibull and finite mixture of the von Mises distribution for wind analysis in Mersing,Malaysia

Studies of wind direction receive less attention than that of wind speed; however, wind direction affects daily activities such as shipping, the use of bridges, and construction. This research aims to study the effect of wind direction on generating wind power. A finite mixture model of the von Mises distribution and Weibull distribution are used in this paper to represent wind direction and wind speed data, respectively, for Mersing (Malaysia). The suitability of the distribution is examined by the R² determination coefficient. The energy analysis, that is, wind power density, only involves the wind speed, but the wind direction is vital in measuring the dominant direction of wind so that the sensor could optimize wind capture. The result reveals that the estimated wind power density is between 18.2 and 25 W/m², and SSW is the most common wind direction for this data.     

GENERATION OF SYNTHETIC AND MISSING CLIMATIC DATA FOR PUERTO RICO1

ABSTRACT: Climatic data such as temperature, solar radiation, relative humidity, and wind speed have been widely used to estimate evapotranspiration. Moat of the solar radiation data and portions of the relative humidity data are either not available or missing from the records in Puerto Rico. Depending upon the availability and data characteristics of records, three methods (including a regression technique, an averaging of historical data, and a regional average) were used to generate missing data, and a time series analysis was used to synthesize a series of climatic data. The limitations and applicability of each method are discussed. The results showed that the time series analysis method can be successfully used to synthesize a series of monthly solar radiations for several stations. The regression technique and the regional average can be successfully applied to generate missing monthly solar radiation data. The regression technique and the averaging of historical data have been satisfactorily used to interpolate missing monthly relative humidity. The explained variance (R²) varied from 0.68 to 0.88, which are both significant at the 0.05 level of significance.     

PREDICTION OF TWO-YEAR PEAK STREAM-DISCHARGES USING NEURAL NETWORKS1

ABSTRACT: The cascade correlation neural network was used to predict the two-year peak discharge (Q₂) for major regional river basins of the continental United States (US). Watersheds ranged in size by four orders of magnitude. Results of the neural network predictions ranged from correlations of 0.73 for 104 test data in the Souris-Red Rainy river basin to 0.95 for 141 test data in California. These results are improvements over previous multilinear regressions involving more variables that showed correlations ranging from 0.26 to 0.94. Results are presented for neural networks trained and tested on drainage area, average annual precipitation, and mean basin elevation. A neural network trained on regional scale data in the Texas Gulf was comparable to previous estimates of Q₂ by regression. Our research shows Q₂ was difficult to predict for the Souris-Red Rainy, Missouri, and Rio Grande river basins compared to the rest of the US, and acceptable predictions could be made using only mean basin elevation and drainage areas of watersheds.     

四川省地级市生态环境质量评价模型

在建立四川省生态质量评价指标体系的基础上,将径向基函数（RBF）神经网络模型用于四川省18个地级市生态质量评价和区划,实现了评价结果的可视化与直观化。评价结果与四川省生态质量实际状况相对比,表明了RBF网络的正确性和实用性,为实现四川省生态质量的保护提供了一条新途径。     

Estimation of daily sunshine duration using support vector machines

Although sunshine duration (SD) is one of the most frequently measured meteorological parameters, there is a lack of measurements in some parts of the world. Hence, it should be estimated accurately for areas where no reliable measurement is possible. The main objective of this study is to evaluate the potential of support vector machine (SVM) approach for estimating daily SD. For this purpose, three different kernels of SVM, such as linear, polynomial, and radial basis function (RBF), were used. Different combinations of five related meteorological parameters, namely cloud cover, maximum temperature (T_max), minimum temperature (T_min), relative humidity (RH), and wind speed (WS), and one astronomic parameter, day length, were considered as the inputs of the models, and the output was obtained as daily SD. Simulated values of the models were compared with ground measured values, and concluded that the usage of the SVM-RBF estimator with combination of all input attributes produced the best results. The coefficient of determination, root mean square error, and mean absolute error were found to be 0.8435, 1.5105 h, and 1.0771 h, respectively, for the pooled four-year daily data set of 14 stations in Turkey. It was also deduced that accuracy increased as the number of attributes increased and the major contribution to this came from RH as compared with T_max, T_min, and WS. This study has shown that the SVM methodology can be a good alternative for conventional and artificial neural network methods for estimating daily SD.     

Thermal Characteristics of Stormwater Runoff from Asphalt and Sod Surfaces1

Abstract: Urban impervious surfaces absorb and store thermal energy, particularly during warm summer months. During a rainfall/runoff event, thermal energy is transferred from the impervious surface to the runoff, causing it to become warmer. As this higher temperature runoff enters receiving waters, it can be harmful to coldwater habitat. In an urban watershed, impervious asphalt surfaces (roads, parking lots, and driveways) and pervious residential lawns comprise a significant portion of the watershed area. A paired asphalt‐turfgrass sod plot was constructed to compare the thermal runoff characteristics between asphalt and turfgrass sod surfaces, to identify meteorological variables that influence these thermal characteristics, and to evaluate evaporative heat loss for runoff from asphalt surfaces. Rainfall simulations were conducted during the summers of 2004 and 2005 under a range of climatic conditions. Asphalt surface temperatures immediately prior to rainfall simulations averaged 43.6°C and decreased an average of 12.3°C over 60 min as rain cooled the surface. In contrast, presimulation sod surface temperatures averaged only 23.3°C and increased an average of 1.3°C throughout the rainfall events. Heat transferred from the asphalt to the runoff resulted in initial asphalt runoff temperatures averaging 35.0°C that decreased by an average of 4.1°C at the end of the event. Sod runoff temperatures averaged only 25.5°C and remained fairly constant throughout the simulations. Multivariable regression equations were developed to predict (1) average asphalt surface temperature (R² = 0.90) and average asphalt runoff temperature (R² = 0.92) as a function of solar radiation, rain temperature, and wind speed, and (2) average sod surface temperature (R² = 0.85) and average sod runoff temperature (R² = 0.94) as a function of solar radiation, rain temperature, rain intensity, and wind speed. Based on a heat balance analysis, existing evaporation equations developed from studies on lakes were not adequate to predict evaporation from runoff on a heated impervious surface. The combined heat from the asphalt and sod plots was an average of 38% less than the total heat had the total area consisted solely of asphalt.     

前馈神经网络在环境评价和预测中的应用进展

本文针对水环境中复杂的不确定性及非线性关系,在水环境不确定性分析的基础上,详细阐述了以BP网络和RBF网络为代表的前馈神经网络法的基本原理,分析了两种方法的优点。同时,本文对两种方法在水环境影响评价工作中的应用现状进行总结,分析了两种方法的研究发展趋势。     

Neural Network Input Selection for Hydrological Forecasting Affected by Snowmelt1

Abstract: Snowmelt largely affects runoff in watersheds in Nordic countries. Neural networks (NN) are particularly attractive for streamflow forecasting whereas they rely at least on daily streamflow and precipitation observations. The selection of pertinent model inputs is a major concern in NNs implementation. This study investigates performance of auxiliary NN inputs that allow short‐term streamflow forecasting without resorting to a deterministic snowmelt routine. A case study is presented for the Rivière des Anglais watershed (700 km²) located in Southern Québec, Canada. Streamflow (Q), precipitations (rain R and snow S, or total P), temperature (T) and snow lying (A) observations, combined with climatic and snowmelt proxy data, including snowmelt flow (Q_SM) obtained from a deterministic model, were tested. NN implemented with antecedent Q and R produced the largest gains in performance. Introducing increments of A and T to the NNs further improved the performance. Long‐term averages, seasonal data, and Q_SM failed to improve the networks.     

Comparison of ANN,MVR, and SWR models for computing thermal efficiency of a solar still

In this paper, the viability of modeling the instantaneous thermal efficiency (η_ith) of a solar still was determined using meteorological and operational data with an artificial neural network (ANN), multivariate regression (MVR), and stepwise regression (SWR). This study used meteorological and operational variables to hypothesize the effect of solar still performance. In the ANN model, nine variables were used as input parameters: Julian day, ambient temperature, relative humidity, wind speed, solar radiation, feed water temperature, brine water temperature, total dissolved solids of feed water, and total dissolved solids of brine water. The η_ith was represented by one node in the output layer. The same parameters were used in the MVR and SWR models. The advantages and disadvantages were discussed to provide different points of view for the models. The performance evaluation criteria indicated that the ANN model was better than the MVR and SWR models. The mean coefficient of determination for the ANN model was about 13% and14% more accurate than those of the MVR and SWR models, respectively. In addition, the mean root mean square error values of 6.534% and 6.589% for the MVR and SWR models, respectively, were almost double that of the mean values for the ANN model. Although both MVR and SWR models provided similar results, those for the MVR were comparatively better. The relative errors of predicted η_ith values for the ANN model were mostly in the vicinity of ±10%. Consequently, the use of the ANN model is preferred, due to its high precision in predicting η_ith compared to the MVR and SWR models. This study should be extremely beneficial to those coping with the design of solar stills.     

An integrated assessment model of carbon sequestration benefits: a case study of Liping county, China

This research attempts to model the complexity of planting trees to increase China's CO(2) sequestration potential by using a GIS-based integrated assessment (IA) approach. We use the IA model to assess the impact of China's Grain for Green reforestation and afforestation program on farmer and state incomes as well as CO(2) sequestration in Liping County, Guizhou Province. The IA model consists of five sub-models for carbon sequestration, crop income, timber income, Grain for Green, and carbon credits. It also includes a complementary qualitative module for assessing program impacts by gender and ethnicity. Using four scenarios with various assumptions about types of trees planted, crop incomes by township, CO(2) credit prices, state subsidies, methods for estimating carbon sequestered, and harvesting of trees, we find great variation in the impact of the Grain for Green program on incomes and on carbon sequestered over a 48 year period at both the county and township levels.     

Location wise comparison of mixture distributions for assessment of wind power potential: A parametric study

Scientific literature discussed various types of mixture models and models derived from maximum entropy principle using short-term wind speed data for their relative assessment. The literature on suitability of these mixture models for long-term data is rarely available. However, for correct assessment of wind power potential both wind speed and wind direction are equally important. Therefore, in this paper, both wind speed and wind direction are simultaneously analyzed using several types of mixture distribution and compared the same with conventional Weibull distribution. For wind speed and wind power density assessment, the mixture distributions such as Weibull--Weibull distribution, Gamma--Weibull distribution, Truncated Normal--Weibull distribution, Truncated Normal--Normal distribution, proposed Truncated Normal--Gamma distribution and Gamma--Gamma distribution along with MEP-distribution are compared with conventional 2-parameter Weibull distribution. Similarly, for wind direction analysis, the finite mixtures of von-Mises distribution are compared with conventional von-Mises distribution. Judgment criteria include R², RMSE, Kolmogorov--Smirnov test and relative percentage error in wind power density. The sites selected are the three onshore locations of India, viz., Calcutta, Trivandrum, and Ahmedabad. The results show that for wind speed assessment, mixture distribution performs better than the conventional Weibull distribution for analyzing wind power density. However, location wise comparison of all mixture distribution is of prime importance. For wind direction analysis, finite mixture of two von-Mises distributions proved to be a suitable candidate for Indian climatology.     

Case Study Comparing Multiple Irrigated Land Datasets in Arizona and Colorado,USA

While there are currently a number of irrigated land datasets available for the western United States (U.S.), there is uncertainty regarding in how they relate to each other. To help understand the characteristics of available irrigated datasets, we compared (1) the Cropland Data Layer (CDL), (2) Moderate Resolution Imaging Spectroradiometer Irrigated Agriculture Dataset (IAD), (3) Digitized Irrigated Land (DIL), and (4) Consumptive Use for Irrigation (CUI) data in Arizona and Colorado, U.S. These datasets were derived from multiple sources at various spatial resolutions and temporal scales. We found spatial and temporal trends among all of them. The datasets showed decreases in irrigated land area in Arizona during the 2000–2010 time period. The change ranges and ratios were similar in all Arizona datasets. Irrigated land in Colorado decreased in DIL and CUI but increased in IAD and CDL. The agreement within the same type of dataset during different time periods was from 60% to 80% (R² from 0.35 to 0.72) in Arizona and from 50% to 80% (R² from 0.23 to 0.68) in Colorado. DIL had the highest agreement (80%) in both states. The agreement among different datasets acquired at approximately the same time frame ranged from 51% to 63% (R² from 0.14 to 0.31) in Arizona and from 47% to 69% (R² from 0.32 to 0.40) in Colorado. The results from this study support a greater understanding of the multiresolution and multitemporal nature of these datasets for various applications.     

Causes of decreased reference evapotranspiration and pan evaporation in the Jinghe River catchment, northern China

This study focused on the changes of reference evapotranspiration (ET₀) and pan evaporation (ET_pan) to study the impacts of climate change on the hydrological cycle in the Jinghe River catchment. Based on the Penman–Monteith equation, the ET₀ was calculated. The temporal trend and spatial distribution of ET₀ and E_pan measured with a 20-cm pan were examined at the 14 stations during 1957–2005. The effects of meteorological factors on the variation of ET₀ were determined by analyzing the trends in themselves with comparison between original climate and detrended climate scenarios and then their sensitivity to ET₀. Both the ET₀ and E_pan showed remarkable decreasing trends from 1957 to 2005 and their decreasing rate was 40.9 and 17.7 mm per 10 years, respectively. Trend analysis of meteorological factors exhibited that the reduction in ET₀ and ET_pan was principally caused by both significant decreases in wind speed and sunshine hours. Furthermore, the decreasing trend of ET₀ was mainly dominated by the significant decrease in wind speed with high sensitivity, to a less extent, by the decrease in net radiation. Although relative humidity is one of the most sensitive variables, its effect on ET₀ was negligible because of its temporal constancy. The contribution of wind speed reduction to decreased ET₀ has increased from 50 to 76.1%, but net radiation, by contrast, decreased from 50 to 23.9%.     

An Analysis of the Wind Energy Potential of Elazig,Turkey

Abstract

In this study, the wind energy potential of Elazig is statistically analyzed based on hourly measured wind speed data over the five-year period from 1998 to 2002. The probability density distributions are derived from cumulative distribution functions. Two probability density functions are fitted to the measured probability distribution on a yearly basis. The wind energy potential of the location is studied based on the Weibull and Rayleigh distributions. It was found that the numerical values of both Weibull parameters (k and c) for Elazig vary over a wide range. The yearly values of k range from 1.653 to 1.878 with an average value of 1.819, while those of c are in the range of 2.757–2.994 m/s with an average value of 2.824 m/s. In addition, yearly mean wind speed and mean power density of Elazig is found as 2.79 m/s and 38.76 W/m², respectively. The wind speed distributions are represented by Weibull distribution and also by Rayleigh distribution, with a special case of the Weibull distribution for k = 2. As a result, the Rayleigh distribution is found to be suitable to represent the actual probability of wind speed data for Elazig.