多元回归和BP人工神经网络模型对混合厌氧消化产气量的预测研究

在中温且控制pH值条件下,对脂肪类单基质和城市污水厂剩余污泥进行混合厌氧消化试验。基于多元回归原理和BP人工神经网络原理,对其建立产气量预测模型。由实验数据计算得出:两个阶段多元回归模型的预测平均准确率分别为75.69%和79.29%;BP神经网络模型的预测平均准确率为79.05%。通过对比两种模型的预测结果可知,两种模型都有较高的预测准确率,但BP模型的预测准确率更高,更适用于混合厌氧消化产气量预测。     

NETWORK AND SUBWATERSHED PARAMETERS EXTRACTED FROM DIGITAL ELEVATION MODELS: THE BILLS CREEK EXPERIENCE1

ABSTRACT: An automated extraction of channel network and sub-watershed characteristics from digital elevation models (DEM) is performed by model DEDNM. This model can process DEM data of limited vertical resolution representing low relief terrain. Such representations often include ill-defined drainage boundaries and indeterminate flow paths. The application watershed is an 84 km² low relief watershed in southwestern Oklahoma. The standard for validation is the network and subwatershed parameters defined by the blue line method on USGS 7.5–minute topographic maps. Evaluation of the generated and validation networks by visual comparisons shows a high degree of correlation. Comparison of selected network parameters (channel length, slope, drainage density, etc.) and of drainage network composition (bifurcation, length, slope, and area ratios) shows that, on the average, the generated parameters are within 5 percent of those derived from the validation network. The largest discrepancies were found for the channel slope values. The results of this application demonstrate that DEDNM effectively addresses network definition problems often encountered in low relief terrain and that it can generate accurate network and subwatershed parameters under those conditions.     

Modelling and Prediction of Diesel Engine Performance using Relevance Vector Machine

Diesel engines are being increasingly adopted by many car manufacturers today, yet no exact mathematical diesel engine model exists due to its highly nonlinear nature. In the current literature, black-box identification has been widely used for diesel engine modelling and many artificial neural network (ANN) based models have been developed. However, ANN has many drawbacks such as multiple local minima, user burden on selection of optimal network structure, large training data size, and over-fitting risk. To overcome these drawbacks, this article proposes to apply an emerging machine learning technique, relevance vector machine (RVM), to model and predict the diesel engine performance. The property of global optimal solution of RVM allows the model to be trained using only a few experimental data sets. In this study, the inputs of the model are engine speed, load, and cooling water temperature, while the output parameters are the brake-specific fuel consumption and the amount of exhaust emissions like nitrogen oxides and carbon dioxide. Experimental results show that the model accuracy is satisfactory even the training data is scarce. Moreover, the model accuracy is compared with that using typical ANN. Evaluation results also show that RVM is superior to typical ANN approach.     

神经网络在空气污染预报中的应用研究

空气污染预报是一项复杂的系统工程,是当今环境科学研究的热点,国内外已有将神经网络法应用于大气污染预报的研究。本论文以PM2．5为例,采用伦敦市PM2．5的小时平均浓度数据,使用传统的BP神经网络建立预报模型,定量预测伦敦市PM2.5的小时平均浓度,探讨了大气污染预报网络的建模过程中,扩大样本集、去除样本集数据噪声和在输入向量中加入气象变量等因素对建模所产生的影响。最后得出结论,适当的选择样本集、气象变量,有利于提高所建立网络模型的预测精度。     

REGIONAL SCALE MODELING OF SURFACE AND GROUND WATER INTERACTION IN THE SNAKE RIVER BASIN1

ABSTRACT: Changes in irrigation and land use may impact discharge of the Snake River Plain aquifer, which is a major contributor to flow of the Snake River in southern Idaho. The Snake River Basin planning and management model (SRBM) has been expanded to include the spatial distribution and temporal attenuation that occurs as aquifer stresses propagate through the aquifer to the river. The SRBM is a network flow model in which aquifer characteristics have been introduced through a matrix of response functions. The response functions were determined by independently simulating the effect of a unit stress in each cell of a finite difference groundwater flow model on six reaches of the Snake River. Cells were aggregated into 20 aquifer zones and average response functions for each river reach were included in the SRBM. This approach links many of the capabilities of surface and ground water flow models. Evaluation of an artificial recharge scenario approximately reproduced estimates made by direct simulation in a ground water flow model. The example demonstrated that the method can produce reasonable results but interpretation of the results can be biased if the simulation period is not of adequate duration.     

Measurement and assessment of carrying capacity of the environment in Ningbo, China

Carrying Capacity of the Environment (CCE) provides a useful measure of the sustainable development of a region. Approaches that use integrated assessment instead of measurement can lead to misinterpretation of sustainable development because of confusion between Environmental Stress (ES) indexes and CCE indexes, and the selection of over-simple linear plus models. The present paper proposes a comprehensive measurement system for CCE which comprises models of natural resources capacity, environmental assimilative capacity, ecosystem services capacity, and society supporting capacity. The corresponding measurable indexes are designed to assess CCE using a carrying capacity surplus ratio model and a vector of surplus ratio of carrying capacity model. The former aims at direct comparison of ES and CCE based on the values of basic indexes, and the latter uses a Euclidean vector to assess CCE states. The measurement and assessment approaches are applicable to Strategic Environmental Assessment (SEA) and environmental planning and management. A case study is presented for Ningbo, China, whereby all the basic indexes of ECC are measured and the CCE states assessed for 2005 and 2010.     

ON THE MISMATCH BETWEEN DATA AND MODELS OF HYDROLOGIC AND WATER RESOURCE SYSTEMS1

ABSTRACT: Many difficulties exist in the matching of models with data. This paper identifies elements of this problem and discusses considerations involved in model evaluation. The well known multivariate linear regression model is used to illustrate the distinctions between accuracy and precision and between estimation and prediction (because the model is commonly misused.) No amount of additional data will improve the accuracy of a poor model. A high R², while indicative of a good matching between the observed data and model estimates, is a poor criterion for judging adequacy of the model to make good predictions of future events. Model evaluation also includes the problem of introducing secondary data and proxy variables into a model. Secondary data frequently enter, for example, the mass, energy and water budget equations because of difficulties in measuring the primary variables. Proxy variables arise because of a desire to collapse a vector of incomparable values, say, of water quality into a single number. Review of the above issues indicates that model evaluation is a multi-criterion problem, often imbedded in a larger framework where models are intended to meet multiple objectives. The mismatch of models and data has increasing legal and social consequences.     

AI-based runoff simulation based on remote sensing observations: A case study of two river basins in the United States and Canada

Data-driven techniques are used extensively for hydrologic time-series prediction. We created various data-driven models (DDMs) based on machine learning: long short-term memory (LSTM), support vector regression (SVR), extreme learning machines, and an artificial neural network with backpropagation, to define the optimal approach to predicting streamflow time series in the Carson River (California, USA) and Montmorency (Canada) catchments. The moderate resolution imaging spectroradiometer (MODIS) snow-coverage dataset was applied to improve the streamflow estimate. In addition to the DDMs, the conceptual snowmelt runoff model was applied to simulate and forecast daily streamflow. The four main predictor variables, namely snow-coverage (S-C), precipitation (P), maximum temperature (T_max), and minimum temperature (T_min), and their corresponding values for each river basin, were obtained from National Climatic Data Center and National Snow and Ice Data Center to develop the model. The most relevant predictor variable was chosen using the support vector machine-recursive feature elimination feature selection approach. The results show that incorporating the MODIS snow-coverage dataset improves the models' prediction accuracies in the snowmelt-dominated basin. SVR and LSTM exhibited the best performances (root mean square error = 8.63 and 9.80) using monthly and daily snowmelt time series, respectively. In summary, machine learning is a reliable method to forecast runoff as it can be employed in global climate forecasts that require high-volume data processing.     

APPLICATION OF GREY MODEL AND ARTIFICIAL NEURAL NETWORKS TO FLOOD FORECASTING1

The main focus of this study was to compare the Grey model and several artificial neural network (ANN) models for real time flood forecasting, including a comparison of the models for various lead times (ranging from one to six hours). For hydrological applications, the Grey model has the advantage that it can easily be used in forecasting without assuming that forecast storm events exhibit the same stochastic characteristics as the storm events themselves. The major advantage of an ANN in rainfall‐runoff modeling is that there is no requirement for any prior assumptions regarding the processes involved. The Grey model and three ANN models were applied to a 2,509 km² watershed in the Republic of Korea to compare the results for real time flood forecasting with from one to six hours of lead time. The fifth‐order Grey model and the ANN models with the optimal network architectures, represented by ANN1004 (34 input nodes, 21 hidden nodes, and 1 output node), ANN1010 (40 input nodes, 25 hidden nodes, and 1 output node), and ANN1004T (14 input nodes, 21 hidden nodes, and 1 output node), were adopted to evaluate the effects of time lags and differences between area mean and point rainfall. The Grey model and the ANN models, which provided reliable forecasts with one to six hours of lead time, were calibrated and their datasets validated. The results showed that the Grey model and the ANN1010 model achieved the highest level of performance in forecasting runoff for one to six lead hours. The ANN model architectures (ANN1004 and ANN1010) that used point rainfall data performed better than the model that used mean rainfall data (ANN1004T) in the real time forecasting. The selected models thus appear to be a useful tool for flood forecasting in Korea.     

EFFECT OF ORIENTATION OF SPATIALLY DISTRIBUTED CURVE NUMBERS IN RUNOFF CALCULATIONS1

ABSTRACT: The NRCS curve number approach to runoff estimation has traditionally been to average or “lump” spatial variability into a single number for purposes of expediency and simplicity in calculations. In contrast, the weighted runoff curve number approach, which handles each individual pixel within the watershed separately, tends to result in larger estimates of runoff than the lumped approach. This work proposes further enhancements that consider not only spatial variability, but also the orientation of this variability with respect to the flow aggregation pattern of the drainage network. Results show that the proposed enhancements lead to much reduced estimates of runoff production. A revised model that considers overland flow lengths, consistent with existing NRCS concepts is proposed, which leads to only mildly reduced runoff estimates. Although more physically‐based, this revised model, which accounts directly for spatially distributed curve numbers and flow aggregation, leads to essentially the same results as the original, lumped runoff model when applied to three study watersheds. Philosophical issues and implications concerning the appropriateness of attempting to disaggregate lumped models are discussed.     

SPACE-TIME TRADEOFFS ACROSS THE HYDROLOGIC DATA SETS OF COMPETING RAINFALL-RUNOFF MODELS: A PRELIMINARY ANALYSIS1

ABSTRACT: The concept of a space-time tradeoff is extended to the hydrologic data sets of competing rainfall-runoff modeling techniques. Examples are given by comparing the performance of a regression model and a quasi-physically based model using data from an experimental catchment and data synthetically generated. Space-time tradeoffs are demonstrated within the data sets of the two modeling techniques, but not across the competing hydrologic data sets.     

Using Multiple Watershed Models to Predict Water,Nitrogen, and Phosphorus Discharges to the Patuxent Estuary1

Boomer, Kathleen M.B., Donald E. Weller, Thomas E. Jordan, Lewis Linker, Zhi‐Jun Liu, James Reilly, Gary Shenk, and Alexey A. Voinov, 2012. Using Multiple Watershed Models to Predict Water, Nitrogen, and Phosphorus Discharges to the Patuxent Estuary. Journal of the American Water Resources Association (JAWRA) 1‐25. DOI: 10.1111/j.1752‐1688.2012.00689.x Abstract: We analyzed an ensemble of watershed models that predict flow, nitrogen, and phosphorus discharges. The models differed in scope and complexity and used different input data, but all had been applied to evaluate human impacts on discharges to the Patuxent River or to the Chesapeake Bay. We compared predictions to observations of average annual, annual time series, and monthly discharge leaving three basins. No model consistently matched observed discharges better than the others, and predictions differed as much as 150% for every basin. Models that agreed best with the observations in one basin often were among the worst models for another material or basin. Combining model predictions into a model average improved overall reliability in matching observations, and the range of predictions helped describe uncertainty. The model average was not the closest to the observed discharge for every material, basin, and time frame, but the model average had the highest Nash–Sutcliffe performance across all combinations. Consistently poor performance in predicting phosphorus loads suggests that none of the models capture major controls. Differences among model predictions came from differences in model structures, input data, and the time period considered, and also to errors in the observed discharge. Ensemble watershed modeling helped identify research needs and quantify the uncertainties that should be considered when using the models in management decisions.     

The Road to NHDPlus — Advancements in Digital Stream Networks and Associated Catchments

A progression of advancements in Geographic Information Systems techniques for hydrologic network and associated catchment delineation has led to the production of the National Hydrography Dataset Plus (NHDPlus). NHDPlus is a digital stream network for hydrologic modeling with catchments and a suite of related geospatial data. Digital stream networks with associated catchments provide a geospatial framework for linking and integrating water‐related data. Advancements in the development of NHDPlus are expected to continue to improve the capabilities of this national geospatial hydrologic framework. NHDPlus is built upon the medium‐resolution NHD and, like NHD, was developed by the U.S. Environmental Protection Agency and U.S. Geological Survey to support the estimation of streamflow and stream velocity used in fate‐and‐transport modeling. Catchments included with NHDPlus were created by integrating vector information from the NHD and from the Watershed Boundary Dataset with the gridded land surface elevation as represented by the National Elevation Dataset. NHDPlus is an actively used and continually improved dataset. Users recognize the importance of a reliable stream network and associated catchments. The NHDPlus spatial features and associated data tables will continue to be improved to support regional water quality and streamflow models and other user‐defined applications.     

Geostatistical spatiotemporal analysis of air temperature as an aid to delineating thermal stability zones in a potential show cave: implications for environmental management

Air temperature in several galleries of the Covadura System (Sorbas Gypsum Karst, Almería) was measured at monthly intervals over a period of 1 year. The spatial temperature distribution for each month was modeled in a geostatistical framework. The mean trend of the air temperature and the difference between each experimental temperature measurement and this trend were calculated over space and time. Both the trend and residual component were characterized using a geostatistical space-time model. A large spatial trend of the air temperature was found due to the orientation of galleries within the cave system and as a function of the distance from the main cave entrance. Kriging was used for the spatial estimation of the time covariance of the residuals. This enabled the delimitation of the cave into three zones of varying environmental risk in the event of being opened to visits by the public, according to the degree of stability of air temperature over space and time. The influence of human presence on the spatial temperature distribution was assessed using data collected during a year (2000/2001) in pilot galleries opened to the public. An average visit corresponding to August was selected comprising 16 people over a period of 53 min. This average visit influenced the spatial temperature pattern at distances of more than 90 m from the cave entrance, according to the geostatistical model adopted. Within this zone the mean thermal increment generated by human presence was estimated to be 0.26 degrees C. The spatiotemporal mathematical model of the cave air temperature has been revealed as a useful tool for the environmental management of show caves.     

区域社会经济发展与生态环境耦合关系研究模型的比较分析

社会经济发展与环境资源相协调是可持续发展的重要途径,判断当前区域社会经济发展与生态环境间的耦合关系是建立有效协调机制的关键。本文从发展、机理方法和应用实例等方面系统介绍了目前在社会经济发展与生态环境之间耦合关系研究中主要的3类模型:计量经济学模型(环境库兹涅茨曲线)、耦合度定量判断模型(包括灰色关联度分析和数理模型)和系统动力学模型。通过3类模型机理和应用情况的对照分析,总结各自优劣势,并认为具有多情景分析和多方案执行评估能力的神经网络模型将成为未来发展趋势。     

Accuracy improvement of various short-term load forecasting models by a novel and unified statistical data-filtering method

ABSTRACT

Time-series and machine-learning methods are being strongly exploited to improve the accuracy of short-term load forecasting (STLF) results. In developing countries, power consumption behaviors could be suddenly changed by different customers, e.g. industrial customers, residential customers, so the load-demand dataset is often unstable. Therefore, reliability assessment of the load-demand dataset is obviously necessary for STLF models. Hence, this paper proposes a novel and unified statistical data-filtering method with the best confidence interval to eliminate unexpected noises/outliers of the input dataset before performing various short-term load forecasting models. This proposed novel data-filtering method, so-called the data pre-processing method, is also compared to other existing data-filtering methods (e.g. Kalman filter, Density-Based Spatial Clustering of Applications with Noise, Wavelet transform, and Singular Spectrum Analysis). By using an SCADA system?-based database of a typical 22kV distribution network in Vietnam, NYISO database, and PJM-RTO database, case studies of short-term load forecasting have been conducted with a conventional ARIMA model, an ANN forecasting model, an LSTM-RNN model, an LSTM-CNN combined model, a deep auto-encoder (DAE) network, a Wavenet-based model, a Wavenet and LSTM hybrid model, and a Wavelet Neural Network (WNN) model, which are to validate the novel and unified statistical data-filtering method proposed. The achieved numerical results demonstrate which the accuracy of the aforementioned STLF models can be significantly improved due to the proposed statistical data-filtering method with the best confidence interval of the input load dataset. The proposed statistical data-filtering method can considerably outperform the existing data-filtering methods.     

DEVELOPMENT OF LARGE SCALE GRIDDED RIVER NETWORKS FROM VECTOR STREAM DATA1

ABSTRACT: The Network Tracing Method (NTM) has been developed to determine gridded coarse river networks for modeling large hydrologic systems. For a coarse resolution grid, the NTM determines the downstream cell of each cell and the distance along the actual meandering flow paths between them. Unlike previously developed methods, the NTM uses fine resolution vector river networks as the source of information of the flow patterns rather than digital elevation models. The main advantage of using vector river networks as input is that they capture the hydrologic terrain features better than topographic data do, particularly in areas of low topographic relief. The NTM was applied to South America with a grid resolution of 1 degree by 1 degree and to the globe with a resolution of 2.815 degrees by 2.8125 degrees. Overall, the method captured the flow patterns well. Generated digital river networks and drainage divides showed minor disagreement with those obtained from existing maps, and most of them were consistent with the resolution of the coarse river network. The majority of estimated basin areas were also close to documented values. River lengths calculated with the NTM, however, were consistently underpredicted.     

KALMAN FILTER ESTIMATION AND PREDICTION OF DAILY STREAM FLOWS: I. REVIEW,ALGORITHM, AND SIMULATION EXPERIMENTS1

An important class of models, frequently used in hydrology for the forecasting of hydrologic variables one or more time periods ahead, or for the generation of synthetic data sequences, is the class of autoregressive(AR) models. As the AR models belong to the family of linear stochastic difference equations, they have both a deterministic and a stochastic component. The stochastic component is often assumed to have a Gaussian distribution. It is well known that hydrologic observations (e.g., stream flows) are heavily affected by noise. To account explicitly for the observation noise, the linear stochastic difference equation is expressed in state variable form and an observation model is introduced. The discrete Kalman filter algorithm can then be used to obtain estimates of the state variable vector. Typically, in hydrologic systems, model parameters, system noise statistics and measurement noise statistics are unknown, and have to be estimated. In this study an adaptive algorithm is discussed which estimates these quantities simultaneously with the state variables. The performance of the algorithm is evaluated by using simulated data.     

Comparison of the daily global solar radiation from different data sources in Northwest China climate

The solar radiation data are of high importance to the solar energy systems. Conventional methods to obtain the solar radiation data are from weather stations, solar radiation models, commercial software databases, and field measurements. In the present study, a new daily global solar radiation model is proposed, by combining the quadratic function of sunshine fraction and sine function of the day of the year. The solar radiation model calculated data are then compared with China Meteorological Data Sharing System (CMDSS) data, TRNSYS data, and field-measured data in Northwest China climate. It is found that the newly proposed solar radiation model has better performance than the other nine solar radiation models in the literature. The solar radiation model calculated data fit well with the CMDSS annually average data. The TRNSYS data are a bit larger than the CMDSS annually average data in summer half year and a little smaller than those in winter half year. The solar radiation model and the CMDSS annually average data have the best correlation, whereas the TRNSYS data and the field-measured data have the worst correlation. The solar radiation model calculated data have the best correlation with the other three data sources.     

AutoRAPID: A Model for Prompt Streamflow Estimation and Flood Inundation Mapping over Regional to Continental Extents

This article couples two existing models to quickly generate flow and flood‐inundation estimates at high resolutions over large spatial extents for use in emergency response situations. Input data are gridded runoff values from a climate model, which are used by the Routing Application for Parallel computatIon of Discharge (RAPID) model to simulate flow rates within a vector river network. Peak flows in each river reach are then supplied to the AutoRoute model, which produces raster flood inundation maps. The coupled tool (AutoRAPID) is tested for the June 2008 floods in the Midwest and the April‐June 2011 floods in the Mississippi Delta. RAPID was implemented from 2005 to 2014 for the entire Mississippi River Basin (1.2 million river reaches) in approximately 45 min. Discretizing a 230,000‐km² area in the Midwest and a 109,500‐km² area in the Mississippi Delta into thirty‐nine 1° by 1° tiles, AutoRoute simulated a high‐resolution (~10 m) flood inundation map in 20 min for each tile. The hydrographs simulated by RAPID are found to perform better in reaches without influences from unrepresented dams and without backwater effects. Flood inundation maps using the RAPID peak flows vary in accuracy with F‐statistic values between 38.1 and 90.9%. Better performance is observed in regions with more accurate peak flows from RAPID and moderate to high topographic relief.