How to evaluate models: Observed vs. predicted or predicted vs. observed?

A common and simple approach to evaluate models is to regress predicted vs. observed values (or vice versa) and compare slope and intercept parameters against the 1:1 line. However, based on a review of the literature it seems to be no consensus on which variable (predicted or observed) should be placed in each axis. Although some researchers think that it is identical, probably because r² is the same for both regressions, the intercept and the slope of each regression differ and, in turn, may change the result of the model evaluation. We present mathematical evidence showing that the regression of predicted (in the y-axis) vs. observed data (in the x-axis) (PO) to evaluate models is incorrect and should lead to an erroneous estimate of the slope and intercept. In other words, a spurious effect is added to the regression parameters when regressing PO values and comparing them against the 1:1 line. Observed (in the y-axis) vs. predicted (in the x-axis) (OP) regressions should be used instead. We also show in an example from the literature that both approaches produce significantly different results that may change the conclusions of the model evaluation.     

Stability of self-reported favourite places and place attachment over a 10-month period

We investigated the reliability and stability of favourite place selections and evaluations of place attachment over a 10-month period. A simple random sample of 1273 respondents representative of the population of the two largest cities in Finland completed a mailed questionnaire. Reliability data was obtained in a second questionnaire (N = 427). We obtained five main favourite place types which were extensively managed natural settings (mainly urban woodlands), urban green spaces (mostly parks), waterside environments, exercise and activity/hobby areas, and indoor and outdoor urban areas. Sixty-four percent of the respondents in the second survey selected the same main favourite place type as 10 months earlier, which is more than what might be expected by chance. Extensively managed natural settings (mainly urban woodlands) and waterside environments were reselected most often. Indoor and outdoor urban favourite places had the lowest stability. We found evidence of people being more consistent in their attachment to natural favourite places than to urban favourite places. The findings lend further credence and credibility to earlier results concerning self-regulation and restorative experiences in favourite places.     

北京山区生态系统稳定性评价模型初步研究

在建立北京山区生态评价体系的基础上,参考有关资料和北京的实际状况确定山区生态系统稳定性评价的标准和分级,选择适当的方法确定评价因子的层次关系和重要度,建立了综合评价模型。并以北京山区的一典型区为例对评价模型进行检验,结果与实际基本相符。     

不同变形失稳模式下隧道围岩稳定性判据研究的进展

计算机模拟隧道围岩失稳研究的困难在于失稳模型不清楚、失稳判据不明确。本文综合分析了目前国内外应用的隧道稳定性判别方法,总结了不同变形失稳机制下的计算模型及其量化的破坏判据,分析了各种方法的研究进展和存在局限,为围岩稳定性数值模拟提供快速、合理、可操作的判据。     

考虑刚度劣化的剪切带-围岩系统稳定性分析

峰后的刚度劣化现象(渐进损伤)在多种煤、岩石、混凝土、土和砂子的循环加载实验中已被观察到。本文考虑了剪切带及带外弹性体的不同的刚度劣化现象,推导了剪切带(断层)-弹性围岩系统在直剪条件下的应力—变形曲线。剪切带的刚度劣化不改变剪切带的总变形,但使剪切带的弹性变形增加、塑性变形降低。考虑弹性体的刚度劣化现象之后,系统的峰后应力—变形曲线呈现非线性特征,这不同于过去忽略刚度劣化现象的结果。在应变软化过程中,剪切带之外弹性体的刚度劣化,使系统的峰后响应由负斜率(Ⅰ类行为)变为正斜率(Ⅱ类变形行为)。较高的弹性体尺寸、较高的脆性、较小的内部长度、较低的围岩初始弹性模量、较高的抗剪强度、较低的残余剪切应力及较高的残余剪切弹性模量,使系统的峰后响应变得陡峭,甚至发生Ⅱ类行为,使系统的失稳(岩爆、冲击地压、采矿诱发地震或矿震)易于发生。     

Impact of varying soil structure on transport processes in different diagnostic horizons of three soil types

When soil structure varies in different soil types and the horizons of these soil types, it has a significant impact on water flow and contaminant transport in soils. This paper focuses on the effect of soil structure variations on the transport of pesticides in the soil above the water table. Transport of a pesticide (chlorotoluron) initially applied on soil columns taken from various horizons of three different soil types (Haplic Luvisol, Greyic Phaeozem and Haplic Cambisol) was studied using two scenarios of ponding infiltration. The highest infiltration rate and pesticide mobility were observed for the Bt₁ horizon of Haplic Luvisol that exhibited a well-developed prismatic structure. The lowest infiltration rate was measured for the Bw horizon of Haplic Cambisol, which had a poorly developed soil structure and a low fraction of large capillary pores and gravitational pores. Water infiltration rates were reduced during the experiments by a soil structure breakdown, swelling of clay and/or air entrapped in soil samples. The largest soil structure breakdown and infiltration decrease was observed for the Ap horizon of Haplic Luvisol due to the low aggregate stability of the initially well-aggregated soil. Single-porosity and dual-permeability (with matrix and macropore domains) flow models in HYDRUS-1D were used to estimate soil hydraulic parameters via numerical inversion using data from the first infiltration experiment. A fraction of the macropore domain in the dual-permeability model was estimated using the micro-morphological images. Final soil hydraulic parameters determined using the single-porosity and dual-permeability models were subsequently used to optimize solute transport parameters. To improve numerical inversion results, the two-site sorption model was also applied. Although structural changes observed during the experiment affected water flow and solute transport, the dual-permeability model together with the two-site sorption model proved to be able to approximate experimental data.     

降低负荷以提高反硝化颗粒污泥的稳定性

在上流式污泥床反应器（USB）内,接种好氧活性污泥,以甲醇为碳源,NO-3为电子受体,经过40多d培养,得到良好的反硝化颗粒污泥,粒径2～3 mm,MLSS为36 g/L,氮去除速率和COD去除速率分别在0.15 g NO₃-N/(g VSS·d)和0.8 g COD/(g VSS·d)。当负荷提高至6.44 g NO₃-N/(L·d)继续运行1周后,观察到反应器内颗粒污泥出现上浮,浓度降低,颗粒粒径多数在3～5 mm,外观呈乳白色。为恢复反应器稳定运行,当负荷降至3.86 g NO₃-N/(L·d)时,上浮现象减轻,当负荷降至2.57 g NO₃-N/(L·d)时,上浮现象消失,颗粒污泥密度由不稳定时的1.0018 g/cm³提高到1.0126 g/cm³,颗粒粘连现象基本消失,污泥氮去除速率在0.19 g NO₃-N/(g VSS·d),连续运行30 d,系统保持稳定。分析认为,颗粒污泥表面微生物生长速度过快是导致不稳定的主要因素,较长的泥龄有利于系统稳定。     

Can market-based policies accomplish the optimal floodplain management? A gap between static and dynamic models

The purpose of this short article is to set static and dynamic models for optimal floodplain management and to compare policy implications from the models. River floodplains are important multiple resources in that they provide various ecosystem services. It is fundamentally significant to consider environmental externalities that accrue from ecosystem services of natural floodplains. There is an interesting gap between static and dynamic models about policy implications for floodplain management, although they are based on the same assumptions. Essentially, we can derive the same optimal conditions, which imply that the marginal benefits must equal the sum of the marginal costs and the social external costs related to ecosystem services. Thus, we have to internalise the external costs by market-based policies. In this respect, market-based policies seem to be effective in a static model. However, they are not sufficient in the context of a dynamic model because the optimal steady state turns out to be unstable. Based on a dynamic model, we need more coercive regulation policies.     

亚硝酸盐氮测定中的稳定性试验

对亚硝酸盐氮测定的清洁水样及1．00mg／L亚硝酸盐氮标准使用溶液分别进行了稳定性试验,试验表明,清洁水样置冰箱内至少可保存30d;1．00mg／L亚硝酸盐氮标准溶液可保存180d,大大延长了国标方法中规定的24h的保存期,既提高了工作效率,又节约了试剂,在实际工作中具有一定的可行性及推广性。