Review and comparison of methods to study the contribution of variables in artificial neural network models

Convinced by the predictive quality of artificial neural network (ANN) models in ecology, we have turned our interests to their explanatory capacities. Seven methods which can give the relative contribution and/or the contribution profile of the input factors were compared: (i) the ‘PaD’ (for Partial Derivatives) method consists in a calculation of the partial derivatives of the output according to the input variables; (ii) the ‘Weights’ method is a computation using the connection weights; (iii) the ‘Perturb’ method corresponds to a perturbation of the input variables; (iv) the ‘Profile’ method is a successive variation of one input variable while the others are kept constant at a fixed value; (v) the ‘classical stepwise’ method is an observation of the change in the error value when an adding (forward) or an elimination (backward) step of the input variables is operated; (vi) ‘Improved stepwise a’ uses the same principle as the classical stepwise, but the elimination of the input occurs when the network is trained, the connection weights corresponding to the input variable studied is also eliminated; (vii) ‘Improved stepwise b’ involves the network being trained and fixed step by step, one input variable at its mean value to note the consequences on the error. The data tested in this study concerns the prediction of the density of brown trout spawning redds using habitat characteristics. The PaD method was found to be the most useful as it gave the most complete results, followed by the Profile method that gave the contribution profile of the input variables. The Perturb method allowed a good classification of the input parameters as well as the Weights method that has been simplified but these two methods lack stability. Next came the two improved stepwise methods (a and b) that both gave exactly the same result but the contributions were not sufficiently expressed. Finally, the classical stepwise methods gave the poorest results.     

Trophic relationships in the recently impounded Bagré reservoir in Burkina Faso

The trophic dynamics of Bagré reservoir which has been recently impounded in Burkina Faso was based on the data collected during 1997–1998 period using the Ecopath model and software. Total fish biomass is 22.63 t km⁻² and mainly represents trophic levels (TLs) 2 and 3. The trophic food chain is relatively long and the overall transfer efficiency is quite low. Grazing foodweb based on primary producers is prominent in the reservoir ecosystem and detritus plays a less significant role. Seasonal and long-term variations in water quality have significant influences on the lower TLs clearly showing a bottom-up functioning of the ecosystem. Environmental degradations, such as siltation occurring in the lake, suggest possible risks in limiting ecosystem productivity.     

自组织神经网络在有害赤潮预警研究中的应用

根据非线性理论及有害赤潮的特点，探讨应用人工神经网络原理进行赤潮预警预报的方法，提出网络可塑性较强，无监督的ART－KOHONEN网络模型，阐述了其基本原理和算法。将该模型应用于辽宁省营口市鲅鱼圈海域，结果表明：该方法能较理想地解决无训练样本的分类识别问题，具有较高的精度，可用于有害赤潮的预警预报工作。     

More time in the kitchen,less time on the streets: the micropolitics of cultivating an ethic of care in alternative food networks

There is increased attention to alternative food efforts as individuals and groups seek to build stronger local food infrastructures to increase accessibility, transparency, and fairness with how food is grown, produced, and distributed. In considering individuals and families contending with food injustices and insecurities; concerns and questions have surfaced about what it means to privilege the leadership and participation of these communities in alternative food efforts. While there are no linear answers to these questions, this paper explores how one statewide food network in the United States seeks to involve youth contending with the juvenile justice system in a job readiness programme, Youth Kitchen, that interfaces the youth with farmers, chef educators, community organisations, and farmers markets. This paper contends that integrating alternative food and juvenile justice work is a complex terrain that both advances social justice and reproduces existing power asymmetries within alternative food networks. The inclusion of accounts from multiple stakeholders in the local food and juvenile justice system generates a multilayered view that moves away from an either sustainability or social justice rubric to a more process-oriented lens that reveals the strategic dilemmas that alternative food networks encounter. On the one hand, the social landscape of this programme promotes an ethic of care and shared ownership between the staff and participating youth. At the same time, akin to many alternative food networks, neoliberal interests bump against this ethic of care and white privilege seeps into staffing patterns and everyday programmes in ways that reproduce the status quo.     

成渝城市群碳排放空间关联结构演化及影响因素

本文基于社会网络分析法（SNA）以及二次分配程序（QAP）方法,利用成渝城市群2005~2016年面板数据,对成渝城市群碳排放空间关联性及影响因素进行研究.结果表明：①成渝城市群碳排放空间关联性显著,呈现出复杂的网络结构形态,样本期内,网络密度由0.16增长至0.68,关联关系数从38个增长为162个.②重庆、成都、绵阳和南充等城市位于网络的中心地位,发出了较多的关联关系,同时发挥着中介作用.③碳排放空间关联网络被划分成为5个层级,层级结构整体较为稳定,然而第一层级与第二层级存在较为严重的断层现象.④空间距离、人口数量差异以及经济水平差异是碳排放关联性的主要驱动因素.城市间空间距离越近、人口数量与经济水平差异越大,越容易产生碳排放关联关系.     

智慧水务背景下的供水管网漏损控制研究进展

供水管网漏损是国内外供水行业面临的普遍难题,不仅造成大量的水资源浪费,而且带来管网水质二次污染风险.因此,管网漏损控制一直是供水行业的研究热点.尤其是近年来智慧水务的发展,使管网基础信息与运行监测数据不断完善,大大推动了管网漏损控制技术的发展.综述了管网破损预测、管网漏损识别定位、管网漏损控制方案优化、数据质量控制与漏损管理系统等方面的研究进展,并分析了当前研究的特点与不足,以期为管网漏损控制技术的发展提供借鉴.     

基于遗传-神经网络算法的含均匀腐蚀缺陷油气管线爆破压力预测研究*

为提高含均匀腐蚀缺陷油气管线爆破压力的预测精度，保障长输油气管线的安全运行，将遗传算法和BP神经网络相结合，建立含均匀腐蚀缺陷油气管线爆破压力预测的遗传-BP神经网络（GA-BPNNs）模型。采用已有文献实验数据，分析对比该模型与AGA NG-18，ASME B31G，修正B31G，PCORRC，DNV RP-F101和SHELL 92等方法用于X46，X52，X60，X65，X80等材质油气管线含均匀腐蚀缺陷时爆破压力的计算误差。结果表明:GA-BPNNs模型用于含均匀腐蚀缺陷油气管线爆破压力预测时，误差在-7.78%～6.06%之间，预测精度明显高于目前国内外通用规范的计算结果；该模型操作简单，适用范围广，工程实用性好，为含缺陷压力管道爆破压力的预测提供更好的思路和方案。     

Implications of network particle tracking (NPT) for ecological model interpretation

Network particle tracking (NPT), building on the foundation of network environ analysis (NEA), is a new development in the definition of coherence relations within and between connected systems. This paper evaluates three ecosystem models in a comparison of throughflow- and storage-based NEA and NPT. Compartments in models with high indirect effects and Finn cycling showed low correlation of NEA storage and throughflow with particle repeat visits and numbers of particles in compartments at steady state. Conversely, the correlation between NEA and NPT results was high with two models having lower indirect effects and Finn cycling. Analysis of ecological orientors associated with NEA showed NPT to fully support conventional NEA results when the common conditions of donor control and steady state are satisfied. Particle trajectories are recorded in the new concept of a particle “passport”. Ability to track and record particle in-system histories enables views of multiple scales and opens the possibility of making pathway-dependent modeling decisions. NPT may also enable modeling of time, allowing integration of Newtonian, organismal and stochastic modeling perspectives in a single comprehensive analysis.     

Prediction of response of zooplankton biomass to climatic and oceanic changes

This paper examines the long-term variation in zooplankton biomass in response to climatic and oceanic changes, using a neural network as a nonlinear multivariate analysis method. Zooplankton data collected from 1951 to 1990 off the shore of northeastern Japan were analyzed. We considered patterns of the Kuroshio and the Oyashio, sea surface temperature, and meteorological parameters as environmental factors that affect zooplankton biomass. Back propagation neural networks were trained to generate mapping functions between environmental variables and zooplankton biomass. The performance of the network models was tested by varying the numbers of input and hidden units. Changes in zooplankton biomass could be predicted from environmental conditions. The neural network yielded predictions with smaller errors than those of predictions determined by linear multiple regression. The sensitivity analysis of networks was used to extract predictive knowledge. The air pressure, sea surface temperature, and some indices of atmospheric circulation were the primary factors for predictions. The patterns of the Kuroshio and the Oyashio demonstrated different effects among sea areas.