基于GIS的凉水国家级自然保护区林火应急路网规划研究

摘 要：对凉水国家级自然保护区林火应急路网分析,进行优化选线,使其达到《全国森林防火规划（2016-2025）》2025年路网密度要求。运用K-means聚类算法选取道路节点,在考虑地形等环境因素下,运用ArcGIS多因素叠加分析选取整体线路,经过实地考察验证线路可行。建立林区应急道路评价体系。最终提高林区道路密度达到3.22 m/hm2,符合国家2025年期望林区道路密度。区域分割指数中的区域面积均值（``ｓ）和区域面积方差（D）分别降低了45.7%和 94.7%。为凉水保护区应急路网开设提供参考。     

基于GIS和BP神经网络的洪灾损失评估模型的研究

本文在分析已有洪灾损失评估模型的基础上,构建了一种基于GIS和BP神经网络的洪灾损失评估模型,包括选取洪水致灾、地形条件、防洪能力、社会经济等因子;然后以鄱阳湖区为例进行了洪灾损失评估系统应用。结果表明建立的基于GIS和BP神经网络的洪灾损失评估模型具有较好的可行性和实用性。     

Impact of different parts of unpaved forest roads on runoff and sediment yield in a Mediterranean area

Surface runoff and sediment production on unpaved forest roads in a humid Mediterranean mountainous area has been studied using a simple portable rainfall simulator at an intensity of 90 mm h^− 1. Thirty six rainfall simulations were carried out on road plots: on the roadbank (12), on the sidecast fill (12), and on the roadbed (12). On the roadbanks, the steady-state runoff coefficient was 85.9% and runoff flow appeared after 63 s on average. On the sidecast fills, the steady-state runoff coefficient was 58.6% and mean time to runoff was 48 s. Finally, on the roadbeds, the steady-state runoff coefficient was 21.5% and mean time to runoff was 41 s. The highest soil loss rate was found on the roadbanks (486.7 g m^− 2), mainly due to low plant cover, soil texture and rock fragments. The total soil erosion on the roadbanks was 3 and 18 times higher than those from the roadbeds and the sidecast fills, respectively. As a consequence, roadbanks can be considered the main source of sediments on the studied sites, but the function of unpaved forest roads as source points for runoff generation is more important.     

Pipe failure rate prediction in water distribution networks using multivariate adaptive regression splines and random forest techniques

ABSTRACT

This paper presents the results of a comparison between multivariate adaptive regression splines (MARS) and random forest (RF) techniques in pipe failure prediction in two water distribution networks. In this regard, pipe diameter, pipe length, pipe installation depth, pipe age and average hydraulic pressure are considered as input variables. Results show that the RF outperforms the MARS which is found as an accurate pipe failure rate predictor. The proposed models are further evaluated through dividing the data into three parts of lower, medium and higher pipe failure rate values. According to the equations produced by MARS technique, three variables of pipe diameter, pipe age and average hydraulic pressure are distinguished as the most effective variables in predicting pipe failure rate in the first case study. Four variables of pipe diameter, pipe length, pipe age and average hydraulic pressure are determined as the most effective variables in the second case study.     

Tracing sediment loss from eroding farm tracks using a geochemical fingerprinting procedure combining local and genetic algorithm optimisation

Eroding farm tracks represent important spatially distributed features in many agricultural landscapes and there is concern over their role in catchment sediment problems. It is, however, important to place eroding farm tracks in the context of catchment sediment sources more generally, especially since the former afford potential for targeted sediment mitigation. A sediment source tracing procedure was therefore used to assess the importance of eroding farm track surfaces as a contemporary primary suspended sediment source relative to inputs from pasture or cultivated topsoils and channel banks/subsurface sources, in the upper River Piddle catchment (~ 100 km²), in southern England. The study provided a timely opportunity to assess the performance of both local and global (genetic algorithm; GA) optimisation techniques in the sediment geochemistry mass balance modelling used to apportion sources. Over the duration of the study, average median source contributions for individual time-integrated suspended sediment samples collected from three sub-catchments ranged between 1 ± 1 and 19 ± 3% for farm track surfaces, 31 ± 3 and 55 ± 2% for pasture topsoils, 1 ± 1 and 19 ± 1% for cultivated topsoils and 23 ± 2 and 49 ± 1% for channel banks/subsurface sources. Comparison of the local and GA optimisation techniques demonstrated that GA with random initial values improved the minimisation of the objective functions compared to local searching by 0.01-0.04% of 5000 repeat Monte Carlo iterations. GA informed by the outputs of the local optimisation as initial values improved corresponding performance by 0.05-0.20%. These findings increased confidence in the outputs from the local optimisation mass balance modelling, but fingerprint property datasets should be treated on an individual basis. Future sediment source tracing studies should always endeavour to combine local and global search tools to avoid the risk of using localised solutions for source apportionment estimates.     

Modeling oblique load carrying capacity of batter pile groups using neural network,random forest regression and M5 model tree

M5 model tree, random forest regression (RF) and neural network (NN) based modelling approaches were used to predict oblique load carrying capacity of batter pile groups using 247 laboratory experiments with smooth and rough pile groups. Pile length (L), angle of oblique load (α), sand density (ρ), number of batter piles (B), and number of vertical piles (V) as input and oblique load (Q) as output was used. Results suggest improved performance by RF regression for both pile groups. M5 model tree provides simple linear relation which can be used for the prediction of oblique load for field data also. Model developed using RF regression approach with smooth pile group data was found to be in good agreement for rough piles data. NN based approach was found performing equally well with both smooth and rough piles. Sensitivity analysis using all three modelling approaches suggest angle of oblique load (α) and number of batter pile (B) affect the oblique load capacity for both smooth and rough pile groups.     

Prediction of rainfall runoff‐induced sediment load from bare land surfaces by generalized regression neural network and empirical model

Based on three rainfall run‐off‐induced sediment transport data for bare surface experimental plots, the generalized regression neural network (GRNN) and empirical models were developed to predict sediment load. Rainfall intensity, slope, rainfall duration, soil particle median diameter, clay content of the soil, rill density and soil particle mass density constituted the input variables of the models while sediment load was the target output. The GRNN model was trained and tested. The GRNN model was found successful in predicting sediment load. Sensitivity analysis by the GRNN model revealed that slope and rainfall duration were the most sensitive parameters. In addition to the GRNN model, two empirical models were proposed: (1) in the first empirical model, all the input variables were related to the sediment load, and (2) in the second empirical model, only rainfall intensity, slope and rainfall duration were related to the sediment load. The empirical models were calibrated and validated. At the calibration stage, the coefficients and the exponents of the empirical models were obtained using the genetic algorithm optimization method. The validated empirical models were also applied to two more experimental data sets: (1) one data set was from a field experiment, and (2) one set was from a laboratory experiment. The results indicated the success of the empirical models in predicting sediment load from bare land surfaces.     

基于BP网络模型的采水地面沉降时空预测

鉴于本构模型和土体参数确定上的困难,在有效应力原理和随机介质理论的基础上,建立采水区地面沉降时空预测的BP神经网络模型,所建模型具有分布参数模型的特征。运用所建模型对其他地面沉降监测点进行了计算和预测,研究表明,所建立的BP网络模型能准确地反映采水地面沉降的时空规律。     

Comparison of multiple regression analysis using dummy variables and a NARX network model: an example of a heavy metal adsorption process

In the present study, the adsorption characteristics of coal fly ash obtained from the Kangal Power Plant, Turkey and activated fly ash in the planetary ball mill were investigated to remove the heavy metal ions from aqueous solutions. The adsorption capacity was compared for the first time using a multiple regression analysis with dummy variables and a non‐linear auto regressive exogenous (NARX) network model. An equation was obtained for all types of adsorbents or heavy metals using the regression of q_e on the dummy variables. The predictive ability of NARX was found to be better than that of multiple regression using dummy variables. These models can also be successfully implemented on the experimental data to evaluate the adsorption process. In addition, fly ash is a low cost alternative since it is a more economical and environmentally friendly adsorbent and it is abundant in both nature and from waste material from industry.