Evaluation of models to upscale the small scale variability of Cd sorption in a case study

Cadmium sorption, basic soil properties and water retention were jointly analyzed in an acidic sandy podzol under pine forest in the North of Germany. Samples were taken along a 10 m transect at a depth of 0.15 m with a sample-support of 0.15 m. The small-scale Cd sorption variability was upscaled in two steps. Firstly, it was simplified and, secondly, aggregated from the sample to the pedon scale. We evaluated different models to simplify Cd sorption variability at different levels of spatial aggregation. Our evaluation method was the numerical simulation of Cd transport in the topsoil where the variability of Cd sorption is the key input.We described Cd sorption with the Freundlich parameterization and tested three models to simplify its spatial variability. The reference model (model 1) had two and the simplified models only one spatially variable sorption parameter. Model 2 varied the parameter K_f of the Freundlich parameterization and set the exponent constant. Model 3 expressed only the linear variability of sorption. Each sample had a scaling factor that related to a constant sorption reference function. The Freundlich parameter K_f of the third simplification model (model 4), was derived by a local pedotransfer function. Its variability was, therefore, filtered by the available variation of a limited number of basic soil properties.The average sorption was at all aggregation levels not significantly different between the models. However, the corresponding uncertainty was smallest for model 3, intermediate for model 4 and largest for model 2. We evaluated the different sorption variability models with the simulation of Cd transport. The mean Cd concentrations in the topsoil predicted by the different models were statistically not different. However, at all support scales, the uncertainties of the predicted mean Cd concentrations and the RMSE's were smallest when model 3 was used, where the error was about 20% at the sample scale and decreased to below 10% at the pedon scale. Therefore, if measurements of sorption isotherms are available, we recommend to use model 3 to derive the mean sorption behavior with minimal uncertainty.     

Uncertainty analysis in near-surface soil moisture estimation on two typical land-use hillslopes

Purpose

Spatial prediction of near-surface soil moisture content (NSSMC) is necessary for both hydrologic modeling and land use planning. However, uncertainties associated with the prediction are always neglected and lack of quantitative analysis. The objective of this study was to investigate the influences of different sources of uncertainty on NSSMC estimation at two typical hillslopes (i.e., tea garden and forest).

Materials and methods

In this study, stepwise multiple regression models with terrain indices and soil texture were built to spatially estimate NSSMC on two typical land use hillslopes (tea garden and forest) at different dates. The uncertainties due to limited sample sizes used for developing regression models (uncertainty of model parameter), digital elevation model resolutions of 1, 2, 3, 4, and 5 m (uncertainty of terrain indices) and spatial interpolations of soil texture by kriging or cokriging with electromagnetic induction (uncertainty of soil texture), were investigated using bootstrap, resampling, and Latin hypercube sampling techniques, respectively.

Results and discussion

The accuracies of NSSMC predictions were acceptable for both tea garden (the Nash-Sutcliffe efficiency or NSE?=?0.34) and forest hillslopes (NSE?=?0.57). The model parameter uncertainty was more important on tea garden hillslope than on forest hillslope. A significant negative correlation (P?<?0.05) was observed between the model parameter uncertainty and the mean NSSMC of the hillslopes, indicating that the model parameter uncertainty was small when the hillslope was wet. The resolution uncertainty from digital elevation model had a minor effect on NSSMC predictions on both hillslopes. The texture uncertainty was weak on NSSMC estimations on tea garden hillslope. However, it was more important than the model parameter uncertainty on the forest hillslope.

Conclusions

Improving the regression model structure and the hillslope soil texture mapping are critical in the accurate spatial prediction of NSSMC on tea garden and forest hillslopes, respectively. This study presents techniques for analyzing three different uncertainties that can be used to identify the main sources of uncertainties in soil mapping.

     

Accounting for variability within map units when linking a pesticide fate model to soil survey

The aim of this study was to estimate the probability of exceedance (POE) of the USEPA health advisory level for particular pesticides in groundwater beneath citrus groves in southwestern Florida. The approach included bootstrapping to assess the uncertainty of the model output due to the variability of soil input data; a weather generator to provide daily rainfall amounts; daily evapotranspiration by the Blaney-Criddle (FAO) method; and a program written to compute POE values. Bootstrapping enabled us to assess the uncertainty of soils inputs by generating pseudo-profiles of soils from pedon characterization data. These pseudo-profiles were used in Monte Carlo simulations that captured the variance of selected soil parameters within soil taxonomic units. Single-name map units (i.e., consociations) were represented by no fewer than three actual pedon characterization data sets for the named soil and/or closely similar soil(s). In the case of a multiple-name map unit (i.e., soil association or soil complex), no fewer than three actual pedons of named and/or similar soils were used to generate pseudo-profiles for each of the named soils in the map unit. Inputs were linked to the pesticide fate model Chemical Movement in Layered Soil, CMLS (Nofziger and Hornsby) to produce cumulative probability curves showing the fraction of applied pesticide leaching below the 1-m depth. This curve was then used to determine the POE for various soil delineations in the groves within the watershed. Multiple POEs were generated for multiple-named map units; the rule was to select the higher(est) POE value as an estimator of environmental risk for each such map unit. Thematic maps are then created using soil and land use coverages and POE's for various pesticides evaluated.

Our approach suggests that the use of cumulative probability functions and probabilities of exceedance of health standards (USEPA HAL) provides needed integration of the uncertainties associated with input parameters, the significance of which can be easily grasped by decision officials. The use of environmental fate models to predict pesticide behavior in soils throughout a landscape has brought about intense scrutiny of soil attribute data and map unit delineations far beyond that envisioned for the present progressive soil survey. The authors also recognize that uncertainties in model formulation, pesticide fate parameters and toxicity further complicate assessments of this nature.     

Quantifying Uncertainty in Critical Loads: (B) Acidity Mass Balance Critical Loads on a Sensitive Site

This paper reports an uncertainty analysis of critical loads for acid deposition for a site in southern England, using the Steady State Mass Balance Model. The uncertainty bounds, distribution type and correlation structure for each of the 18 input parameters was considered explicitly, and overall uncertainty estimated by Monte Carlo methods. Estimates of deposition uncertainty were made from measured data and an atmospheric dispersion model, and hence the uncertainty in exceedance could also be calculated. The uncertainties of the calculated critical loads were generally much lower than those of the input parameters due to a “compensation of errors” mechanism – coefficients of variation ranged from 13% for CL_maxN to 37% for CL(A). With 1990 deposition, the probability that the critical load was exceeded was > 0.99; to reduce this probability to 0.50, a 63% reduction in deposition is required; to 0.05, an 82% reduction. With 1997 deposition, which was lower than that in 1990, exceedance probabilities declined and uncertainties in exceedance narrowed as deposition uncertainty had less effect. The parameters contributing most to the uncertainty in critical loads were weathering rates, base cation uptake rates, and choice of critical chemical value, indicating possible research priorities. However, the different critical load parameters were to some extent sensitive to different input parameters. The application of such probabilistic results to environmental regulation is discussed.     

Modelling spatial uncertainty of soil erodibility factor using joint stochastic simulation

Soil erosion varies greatly over space and is commonly estimated using the revised universal soil loss equation (RUSLE). Neglecting information about estimation uncertainty, however, may lead to improper decision‐making. One geostatistical approach to spatial analysis is joint stochastic simulation, which draws alternative, equally probable, joint realizations of a regionalized variable. Differences between the realizations provide a measure of spatial uncertainty and allow us to carry out an error propagation analysis. The objective of this paper was to assess spatial uncertainty of a soil erodibility factor (K) model resulting from the uncertainties in the input parameters (texture and organic matter). The 500 km² study area was located in central‐eastern Sardinia (Italy) and 152 samples were collected. A Monte Carlo analysis was performed where spatial cross‐correlation information through joint turning bands simulation was incorporated. A linear coregionalization model was fitted to all direct and cross‐variograms of the input variables, which included three different structures: a nugget effect, a spherical structure with a shorter range (3500 m) and a spherical structure with a longer range (10 000 m). The K factor was then estimated for each set of the 500 joint realizations of the input variables, and the ensemble of the model outputs was used to infer the soil erodibility probability distribution function. This approach permitted delineation of the areas characterized by greater uncertainty, to improve supplementary sampling strategies and K value predictions. Copyright © 2007 John Wiley & Sons, Ltd.     

Uncertainty analysis for pedotransfer functions

Both empirical and process‐simulation models are useful in predicting the outcome of agricultural management on soil quality and vice versa, and pedotransfer functions have been developed to translate readily available soil information into variables that are needed in the models. The input data are subject to error, and consequently the transfer functions can produce varied outputs. A general approach to quantifying the resulting uncertainty is to use Monte Carlo methods. By sampling repeatedly from the assumed probability distributions of the input variables and evaluating the response of the model, the statistical distribution of the outputs can be estimated. Methods for sampling the probability distribution include simple random sampling, the sectioning method, and Latin hypercube sampling. The Latin hypercube sampling is applied to the quantification of uncertainties in pedotransfer functions of soil strength and soil hydraulic properties. Hydraulic properties predicted using recently developed pedotransfer functions are also used in a model to analyse the uncertainties in the prediction of soil‐water regimes in the field. The uncertainties of hydraulic properties in soil‐water simulation show that the model is sensitive to the soil's moisture state.     

Predicting regional emissions and near-field air concentrations of soil fumigants using modest numerical algorithms: a case study using 1,3-dichloropropene

Soil fumigants, used to control nematodes and crop disease, can volatilize from the soil application zone and into the atmosphere to create the potential for human inhalation exposure. An objective for this work is to illustrate the ability of simple numerical models to correctly predict pesticide volatilization rates from agricultural fields and to expand emission predictions to nearby air concentrations for use in the exposure component of a risk assessment. This work focuses on a numerical system using two U.S. EPA models (PRZM3 and ISCST3) to predict regional volatilization and nearby air concentrations for the soil fumigant 1,3-dichloropropene. New approaches deal with links to regional databases, seamless coupling of emission and dispersion models, incorporation of Monte Carlo sampling techniques to account for parametric uncertainty, and model input sensitivity analysis. Predicted volatility flux profiles of 1,3-dichloropropene (1,3-D) from soil for tarped and untarped fields were compared against field data and used as source terms for ISCST3. PRZM3 can successfully estimate correct order of magnitude regional soil volatilization losses of 1,3-D when representative regional input parameters are used (soil, weather, chemical, and management practices). Estimated 1,3-D emission losses and resulting air concentrations were investigated for five geographically diverse regions. Air concentrations (15-day averages) are compared with the current U.S. EPA's criteria for human exposure and risk assessment to determine appropriate setback distances from treated fields. Sensitive input parameters for volatility losses were functions of the region being simulated.     

孔隙结构对水稻土温室气体排放的影响

土壤结构影响水分和气体的运动和土壤生物活动,进而影响稻田温室气体排放.为探明土壤结构对水稻生长过程中温室气体排放的影响,选取江苏宜兴的湖白土和江西进贤的红壤性水稻土进行盆栽试验.设置不搅动(NP)、搅动(PD)和搅动后掰土回填(RP)3个处理.应用X射线CT成像技术分析不同处理土壤孔隙结构,通过静态箱法测定水稻生长过程...     

非均匀土壤区域SVAT模型中土壤水动力学参数的整合研究

荷兰Cabauw地区一个 10km×10km区域内的四种主要土壤的水动力学参数用一种实验室直接测定法 (Wind氏蒸发法 )和两种间接方法 (“分段土壤推导函数”、“连续土壤推导函数”)确定 ,而该区域水动力学特性参数的整合则采用两组方案 (聚合土壤参数法和有效土壤参数法 )进行。一个SVAT模型的模拟输出结果———感热通量、潜热通量与实测数据的比较分析表明 :(1)对于用“分段土壤推导函数”确定的土壤水动力学参数的区域整合 ,以“逆模拟法”(有效土壤参数法 )较为可行 ,其模拟感热、潜热的精确性接近参比方案 (“模拟 平均法”) ;(2 )对于实验室直接测定的参数 ,则以几何平均vanGenuchten Mualem经验公式参数的方案 (聚合土壤参数法 )为佳 ;(3)对于“连续土壤推导函数”推导的水动力学参数 ,几何平均土壤组分方案 (聚合土壤参数法 )和“逆模拟法”方案 (有效土壤参数法 )二者均可得到优于参比方案 (“模拟 平均法”)的模拟结果 ,其中以前者最佳 ;(4 )所有区域化参数整合方案中 ,以水平几何平均区域内实验室直接测定的参数的方案最优 ;同时 ,“连续土壤推导函数”法的土壤组分几何平均方案的模型输出精确性接近该最优方案     

非均匀土壤区域SVAT模型中土壤水动力学参数的整合研究

荷兰Cabauw地区一个 10km10km区域内的四种主要土壤的水动力学参数用一种实验室直接测定法 (Wind氏蒸发法 )和两种间接方法 (分段土壤推导函数、连续土壤推导函数)确定 ,而该区域水动力学特性参数的整合则采用两组方案 (聚合土壤参数法和有效土壤参数法 )进行。一个SVAT模型的模拟输出结果感热通量、潜热通量与实测数据的比较分析表明 :(1)对于用分段土壤推导函数确定的土壤水动力学参数的区域整合 ,以逆模拟法(有效土壤参数法 )较为可行 ,其模拟感热、潜热的精确性接近参比方案 (模拟 平均法) ;(2 )对于实验室直接测定的参数 ,则以几何平均vanGenuchten Mualem经验公式参数的方案 (聚合土壤参数法 )为佳 ;(3)对于连续土壤推导函数推导的水动力学参数 ,几何平均土壤组分方案 (聚合土壤参数法 )和逆模拟法方案 (有效土壤参数法 )二者均可得到优于参比方案 (模拟 平均法)的模拟结果 ,其中以前者最佳 ;(4 )所有区域化参数整合方案中 ,以水平几何平均区域内实验室直接测定的参数的方案最优 ;同时 ,连续土壤推导函数法的土壤组分几何平均方案的模型输出精确性接近该最优方案     

Quantifying Uncertainty in Critical Loads: (A) Literature Review

Critical loads are the basis for policies controlling emissions of acidic substances in Europe. The implementation of these policies involves large expenditures, and it is reasonable for policymakers to ask what degree of certainty can be attached to the underlying critical load and exceedance estimates. This paper is a literature review of studies which attempt to estimate the uncertainty attached to critical loads. Critical load models and uncertainty analysis are briefly outlined. Most studies have used Monte Carlo analysis of some form to investigate the propagation of uncertainties in the definition of the input parameters through to uncertainties in critical loads. Though the input parameters are often poorly known, the critical load uncertainties are typically surprisingly small because of a “compensation of errors” mechanism. These results depend on the quality of the uncertainty estimates of the input parameters, and a “pedigree” classification for these is proposed. Sensitivity analysis shows that some input parameters are more important in influencing critical load uncertainty than others, but there have not been enough studies to form a general picture. Methods used for dealing with spatial variation are briefly discussed. Application of alternative models to the same site or modifications of existing models can lead to widely differing critical loads, indicating that research into the underlying science needs to continue.     

A Bayesian framework for model calibration, comparison and analysis: Application to four models for the biogeochemistry of a Norway spruce forest

Four different parameter-rich process-based models of forest biogeochemistry were analysed in a Bayesian framework consisting of three operations: (1) Model calibration, (2) Model comparison, (3) Analysis of model-data mismatch.Data were available for four output variables common to the models: soil water content and emissions of N₂O, NO and CO₂. All datasets consisted of time series of daily measurements. Monthly averages and quantiles of the annual frequency distributions of daily emission rates were calculated for comparison with equivalent model outputs. This use of the data at model-appropriate temporal scale, together with the choice of heavy-tailed likelihood functions that accounted for data uncertainty through random and systematic errors, helped prevent asymptotic collapse of the parameter distributions in the calibration.Model behaviour and how it was affected by calibration was analysed by quantifying the normalised RMSE and r² for the different output variables, and by decomposition of the MSE into contributions from bias, phase shift and variance error. The simplest model, BASFOR, seemed to underestimate the temporal variance of nitrogenous emissions even after calibration. The model of intermediate complexity, DAYCENT, simulated the time series well but with large phase shift. COUP and MoBiLE-DNDC were able to remove most bias through calibration.The Bayesian framework was shown to be effective in improving the parameterisation of the models, quantifying the uncertainties in parameters and outputs, and evaluating the different models. The analysis showed that there remain patterns in the data - in particular infrequent events of very high nitrogenous emission rate - that are unexplained by any of the selected forest models and that this is unlikely to be due to incorrect model parameterisation.     

Application of the fast method to analyze the sensitivity-uncertainty of a lagrangian model of sulphur transport in Europe

A Fourier Amplitude Sensitivity Test (FAST) is applied to study uncertainty of the EMEP-W atmospheric model of long range transport of S in Europe. The FAST method requires frequency distribution of model parameters as input data and provides the following results: (i) mean value of the model output, (ii) variance of the model output which characterizes the parameter uncertainty of the model, and (iii) partial variances of the model output which are the measures of the model sensitivity to uncertainties in individual input parameters. A mathematical formulation of the FAST method and approximations used in its computer implementation is presented. The application requires an extension of the original method to evaluate the model output frequency distribution for different forms of prescribed frequency distributions for parameters. The computer program allows one to apply the FAST method for an arbitrary mathematical model with different options defined by the user. Some examples of uncertainty analysis of the EMEP model using real meteorological and emission data are described and compared with results obtained by the Monte-Carlo method.     

An assessment of factors controlling N<Subscript>2</Subscript>O and CO<Subscript>2</Subscript> emissions from crop residues using different measurement approaches

Management of plant residues plays an important role in maintaining soil quality and nutrient availability for plants and microbes. However, there is considerable uncertainty regarding the factors controlling residue decomposition and their effects on greenhouse gas (GHG) emissions from the soil. This uncertainty is created both by the complexity of the processes involved and limitations in the methodologies commonly used to quantify GHG emissions. We therefore investigated the addition of two soil residues (durum wheat and faba bean) with similar C/N ratios but contrasting fibres, lignin and cellulose contents on nutrient dynamics and GHG emission from two contrasting soils: a low-soil organic carbon (SOC), high pH clay soil (Chromic Haploxerert) and a high-SOC, low pH sandy-loam soil (Eutric Cambisol). In addition, we compared the effectiveness of the use of an infrared gas analyser (IRGA) and a photoacoustic gas analyser (PGA) to measure GHG emissions with more conventional gas chromatography (GC). There was a strong correlation between the different measurement techniques which strengthens the case for the use of continuous measurement approaches involving IRGA and PGA analyses in studies of this type. The unamended Cambisol released 286% more CO₂ and 30% more N₂O than the Haploxerert. Addition of plant residues increased CO₂ emissions more in the Haploxerert than Cambisol and N₂O emission more in the Cambisol than in the Haploxerert. This may have been a consequence of the high N stabilization efficiency of the Haploxerert resulting from its high pH and the effect of the clay on mineralization of native organic matter. These results have implication management of plant residues in different soil types.     

Characteristics and issues related to regional-scale modeling of nitrogen flows

This paper summarizes the characteristics of regional-scale nitrogen (N) flow models. The regional scale is generally considered to be an area that ranges from more than 10 km² to the size of a continent. Parameterization is the key process in creating a regional-scale model. During parameterization, transfer functions that reflect the controlling factors must be created at the target scale because the influence of different factors will change with the size of the scale. Watersheds are the most useful unit for evaluating overall N discharge; however, regional activity data is most often available for municipal units. Thus, municipal units must be reaggregated into watershed units. A longer time period is desirable to normalize seasonal and annual variations at regional scales. Parameters that influence N flow must match the investigated spatial and temporal scales. Given the need to use a range of parameters that vary in terms of the quality of the data, models exhibit inevitable uncertainties. Quantification of the uncertainties and verification of the estimated results are required. Error propagation, the Monte Carlo simulation method and maximum and minimum values have been used to obtain different threshold values of uncertainty. To verify regional-scale N flow models, the following five approaches have been used or proposed: (1) calibration of the model by detailed monitoring at multiple sites, (2) verification of the most important process of the extrapolation mechanisms, (3) verification of the N budget, paying particular attention to water quality, (4) comparison with the results quantified by different models, (5) comparison with aerial or satellite image analysis. As regional-scale modeling of N flow will become more important in the future, it is important to develop models than can accurately estimate N dynamics at this scale.     

Uncertainties in up-scaling N2O flux from field to 1°×1° scale: A case study for Inner Mongolian grasslands in China

Since N₂O emissions cannot be measured easily at large scales, global emission estimates inevitably involve problems with scaling. To date, up-scaling processes depend highly on the models and database. Because of the limitation in resolution of the databases, which provide input parameters to drive the model's regional simulations, the uncertainties generated from the up-scaling processes must be quantified. In this paper, the uncertainties in up-scaling N₂O emissions from the field scale (∼1 km²) to 1°×1° scale (∼10,000 km²) were quantified in a case study from the Xilin River basin of Inner Mongolia, China. A revised process-based DNDC model was applied in the study for quantifying N₂O fluxes with a high-resolution (1 km²) soil database constructed with remote sensing data and GIS technique. The results showed that the uncertainties coming from spatial scaling effect is 63.6%, and from the partitioning of sensitive model parameter (SOC) is 86.4%. We found that inclusion of spatial heterogeneity of soil factors resulted in lower regional N₂O emission estimates. Utilization of the spatial structural information based on soil type was more effective for reducing the spatial scaling effect in comparison with the variability information calculated from Monte Carlo method.     

华北冬小麦/夏玉米农田水氮管理的温室效应评价

【目的】农田水肥的高投入在保障粮食产量的同时也伴随着温室气体的排放。本研究以农田投入品的生产和运输—作物生长的整个过程为研究对象,对农田生态系统中不同水氮管理措施的温室效应开展了评价。【方法】在已经确定粮食产量和温室气体排放强度评价指标的基础上,对土壤固碳深度和减排措施评价的时间尺度进行了分析,将土壤固碳深度确定为30 cm以上,温室效应评价的时间尺度确定为20 年,提出了以田间试验与过程模型相结合,辅以调研的评价方法,来反映产量和温室气体排放强度对不同管理措施的响应。常规农民措施的农田投入量(灌溉量和施氮量等)通过问卷调研和文献数据来获得。以华北平原冬小麦/夏玉米轮作模式为例,利用验证后的农田生态系统管理模型(APSIM)对不同措施(氮肥利用、 灌溉和有机肥配施)进行20 年 (1990~2010)尺度的模拟,并结合农田投入品在生产和运输过程的排放,遵照评价方法对不同水氮管理措施的温室效应进行了分析。【结果】与当前常规农民措施相比,将常规施氮量从520 kg/hm2减少为400 kg/hm2,粮食产量在20 a间不存在显著性差异(P=0.39),但年均温室气体排放总量(AE-GHG)可减少约1.45 t/hm2,温室气体排放强度(GHGI)可减少约0.08 t/t; 若将该地区常规灌溉量从300 mm减少到240 mm,粮食产量在20年间不存在显著性差异(P=0.39),年均温室气体排放总量(AE-GHG)可减少约0.29 t/hm2,温室气体排放强度(GHGI)可减少约0.01 t/t,主要归因于电力消耗的降低,减少了生产和传输过程中温室气体排放; 若将常规措施中的底肥(N)全部替换为有机肥,粮食产量在20年间不存在显著性差异(P=0.63),年均温室气体排放总量(AE-GHG)可减少约0.03 t/hm2,温室气体排放强度(GHGI)则基本无变化,虽然有机肥可带来更多的土壤固碳,但是若考虑到其堆肥生产排放和还田过程增加的油耗排放,其总体温室气体减排量并不明显。【结论】在华北平原当前情况下,农田温室气体减排措施应以减施化肥、 减少灌溉量为主要方向,可同时实现氮肥生产运输和农田土壤排放2个环节上的减排。年施用氮肥减少120 kg/hm2,灌溉量减少60 mm,20年内温室气体减排潜力约为1.45和0.29 t/(hm2·a)。     

Climate Projections Using Bayesian Model Averaging and Space–Time Dependence

Projections of future climatic changes are a key input to the design of climate change mitigation and adaptation strategies. Current climate change projections are deeply uncertain. This uncertainty stems from several factors, including parametric and structural uncertainties. One common approach to characterize and, if possible, reduce these uncertainties is to confront (calibrate in a broad sense) the models with historical observations. Here, we analyze the problem of combining multiple climate models using Bayesian Model Averaging (BMA) to derive future projections and quantify uncertainty estimates of spatiotemporally resolved temperature hindcasts and projections. One advantage of the BMA approach is that it allows the assessment of the predictive skill of a model using the training data, which can help identify the better models and discard poor models. Previous BMA approaches have broken important new ground, but often neglected space–time dependencies and/or imposed prohibitive computational demands. Here we improve on the current state-of-the-art by incorporating space–time dependence while using historical data to estimate model weights. We achieve computational efficiency using a kernel mixing approach for representing a space–time process. One key advantage of our new approach is that it enables us to incorporate multiple sources of uncertainty and biases, while remaining computationally tractable for large data sets. We introduce and apply our approach using BMA to an ensemble of Global Circulation Model output from the Intergovernmental Panel on Climate Change Fourth Assessment Report of surface temperature on a grid of space–time locations.     

农业土壤氧化亚氮排放模型研究进展

农业土壤产生的氧化亚氮气体（N2O）是重要人为N2O源。农业土壤N2O排放模型众多,根据模型建立方法的不同,可分为过程机理模型和经验模型。为探讨产生N2O的具体过程（硝化过程和反硝化过程）和关键因子,着重介绍了DNDC、DAYCENT、Ecosys、WNMM等机理过程模型,指出尽管各个模型的N循环过程类似,但不同侧重因子造成N2O排放量不同,并列出不同模型的特点和应用现状。对目前应用得比较广泛的经验统计模型,如经验归纳模型、回归模型以及其他统计模型等,归纳了其特点并介绍了国内外研究进展。通过对比过程机理模型和经验统计模型的优缺点,指出前者参数较多、过程复杂,用于点位模拟准确度高,后者所需参数少,适用区域范围模拟,点位模拟结果不确定性差。在此基础上指出区域N2O模拟排放量和排放特性将是以后发展的重点方向,并提出区域模拟关键问题的解决方向。     

Modeling of maximum dry density and optimum moisture content of stabilized soil using artificial neural networks

This study considers the use of artificial neural networks (ANNs) to predict the maximum dry density (MDD) and optimum moisture content (OMC) of soil‐stabilizer mix. Multilayer perceptron (MLP), one of the most widely used ANN architectures in the literature, is utilized to construct comprehensive and accurate models relating the MDD and OMC of stabilized soil to the properties of natural soil such as particle‐size distribution, plasticity, linear shrinkage, and the type and quantity of stabilizing additives. Five ANN models are constructed using different combinations of the input parameters. Two separate sets of ANN prediction models, one for MDD and the other for OMC, and also a combined ANN model for multiple outputs are developed using the potentially influential input parameters. Relative‐importance values of various inputs of the models are calculated to determine the significance of each of the predictor variables to MDD and OMC. Inferring the most relevant input parameters based on Garson's algorithm, modified ANN models are separately developed for MDD and OMC. The modified ANN models are utilized to introduce explicit formulations of MDD and OMC. A parametric study is also conducted to evaluate the sensitivity of MDD and OMC due to the variation of the most influencing input parameters. A comprehensive set of data including a wide range of soil types obtained from the previously published stabilization test results is used for training and testing the prediction models. The performance of ANN‐based models is subsequently analyzed and compared in detail. The results demonstrate that the accuracy of the proposed models is satisfactory as compared to the experimental results.