Model averaging for genome‐enabled prediction with reproducing kernel Hilbert spaces: a case study with pig litter size and wheat yield

Predictive ability of yet‐to‐be observed litter size (pig) grain yield (wheat) records of several reproducing kernel Hilbert spaces (RKHS) regression models combining different number of Gaussian or t kernels was evaluated. Predictive performance was assessed as the average (over 50 replicates) predictive correlation in the testing set. Predictions from these models were combined using three different types of model averaging: (i) mean of predicted phenotypes obtained in each model, (ii) weighted average using mean squared error as weight or (iii) using the marginal likelihood as weight. (ii) and (iii) were obtained in a validation set with 5% of the data. Phenotypes consisted of 2598, 1604 and 1879 average litter size records from three commercial pig lines and wheat grain yield of 599 lines evaluated in four macro‐environments. SNPs from the PorcineSNP60 BeadChip and 1447 DArT markers were used as predictors for the pig and wheat data analyses, respectively. Gaussian and univariate t kernels led to same predictive performance. Multikernel RKHS regression models overcame shortcomings of single kernel models (increasing the predictive correlation of RKHS models by 0.05 where 3 Gaussian or t kernels were fitted in the RKHS models simultaneously). None of the proposed averaging strategies improved the predictive correlations attained with single models using multiple kernel fitting.     

非线性二阶阻尼微分方程的区间振动准则

给出非线性二阶阻尼微分方程新的振动准则,它们仅依赖于区间[t0,∞)的某些子区间序列的信息,不需要依赖于整个半直线,所得结果推广和改进了某些最近的结果.     

饱和修正系数提高ZZLAS型闪烁仪测量显热通量精度

为了研究饱和效应对大孔径闪烁仪估算区域显热通量的影响,在2014年8—9月期间进行试验,以具有抗饱和性能的双光路BLS900型闪烁仪为参考,以孔径尺寸为0.075 m(文中简记为LAS1)、0.15 m(文中简记为LAS2)的中国产ZZLAS型闪烁仪为研究对象,通过光学传播原理计算出饱和修正系数,并对ZZLAS型闪烁仪的观测结果进行饱和修正分析。结果表明:LAS1的饱和修正系数为1.034,LAS2的饱和修正系数为1.019。试验观测中LAS1饱和率为24.58%,LAS2饱和率为2.04%,进行饱和修正后,LAS1的有效饱和修正率为12.87%。与BLS900相比,LAS1修正后显热通量均方根误差变为25.67 W/m2;LAS2的饱和修正率仅为0.32%,修正前后显热通量均方根误差基本无变化。进行饱和修正前,对BLS900与LAS1观测的显热通量进行F检验,未达到显著水平(P=0.15);通过计算得出的修正系数修正后,达到极显著水平(P=0.004);而利用BLS900的实时修正系数进行修正后,接近显著水平(P=0.06)。利用试验计算得出的饱和修正系数修正后,显热通量与参考标准的误差范围为1.28~53.42 W/m2,比修正前更接近BLS900的观测结果。研究对农田、人工林下垫面条件下的观测结果采用文中的饱和修正方法进行验证,结果也表明,经饱和修正后,闪烁仪观测的显热通量更接近BLS900的观测结果。当ZZLAS型闪烁仪发生明显饱和现象时,利用光学传播原理计算得出的修正系数对饱和数据进行修正效果显著。     

Interpretation of lysimeter weighing data affected by wind

Weighing lysimeters are valuable devices for measuring water‐balance components with high temporal resolution and high accuracy. However, some older lysimeter facilities still operate with lever‐arm‐counterbalance weighing systems that are sensitive to disturbances, e.g., forces exerted by wind. Filtering and averaging are commonly used for processing noisy raw data. We studied some data of a lever‐arm weighing system and performed additional experiments in order to (1) determine the measurement accuracy of the current weighing scheme (facility, and measuring and averaging procedure) regarding wind effects, (2) describe the oscillation behavior, (3) test the mechanical performance of the system, and (4) adapt the averaging procedure with respect to improved interpretation of the weighing data. The measurement accuracy for a wind velocity < 5 m s^–1, measured in 10 m height, was ≈ ± 0.4 kg (equivalent to ± 0.14 mm); at a higher wind velocity, the accuracy was three times lower, but there was no linear relationship. Additional experiments showed that the weighing system is oscillating with more or less irregular amplitudes. A loading–unloading experiment delivered proper results of the measured loads. The mechanical system reacted immediately, and no directional effects were found. However, small changes of < 1 kg could hardly be determined due to the oscillations. A time series of raw data measured every 2 s served as basis for improving the averaging method. A moving average from 64 values was computed representing the currently used method, and serving as reference. With this procedure an accuracy of ± 0.38 kg could be reached. Averaging 150 values led to an accuracy of ± 0.28 kg (0.1 mm) for a wind velocity < 5 m s^–1.     

Detrimental effects of peat mining on amphibian abundance and species richness in bogs

Peat bogs of eastern North America are increasingly being mined for peat, which results in the extensive draining of these systems. First, I quantified the effects of peat mining on amphibian species richness and abundance in bog remnants adjacent to mining activity and unmined bogs. I then investigated amphibian occurrence patterns in bog remnants relative to the distance to the mined edge, after accounting for microhabitat and distance to landscape elements. The response of amphibians to peat mining varied across taxa. For instance, amphibian species richness, total amphibian captures (all spp. combined), as well as those of green frogs (Rana clamitans) were greater in unmined bogs than in bog remnants. Wood frogs (Rana sylvatica) were most abundant in unmined bogs in areas far from bog ponds, as well as in unmined bogs in areas near shrub patches. After accounting for microhabitat and landscape variables, American toads (Bufo americanus) in bog remnants were most abundant close to mined edges, whereas the other species did not respond to the proximity of mined edges. Peat mining negatively impacts amphibians occurring in bog remnants. The response of amphibians to peat mining is similar to the one of birds and plants, but differs with small mammals. Maintaining a complex mosaic of bog ponds, shrub and forest patches might help mitigate peat mining effects on amphibian patterns of abundance.     

Sugar maple (Acer saccharum Marsh.) growth is influenced by close conspecifics and skid trail proximity following selection harvest

In this study, we quantified the effects of local neighbourhood competition, light availability, and proximity to skid trails on the growth of sugar maple (Acer saccharum Marsh.) trees following selection harvest. We hypothesized that growth would increase with decreasing competition and increasing light availability, but that proximity to skid trails would negatively affect growth. A total of 300 sugar maples were sampled ∼10 years after selection harvesting in 18 stands in Témiscamingue (Québec, Canada). Detailed tree and skid trail maps were obtained in one 0.4 ha plot per stand. Square-root transformed radial growth data were fitted to a linear mixed model that included tree diameter, crown position, a neighbourhood competition index, light availability (estimated using the SORTIE light model), and distance to the nearest skid trail as explanatory variables. We considered various distance-dependent or -independent indices based on neighbourhood radii ranging from 6 to 12 m. The competition index that provided the best fit to the data was a distance-dependent index computed in a 6 m search radius, but a distance-independent version of the competition index provided an almost equivalent fit to data. Models corresponding to all combinations of main effects were fit to data using maximum likelihood, and weighted averages of parameter estimates were obtained using multimodel inference. All predictors had an influence on growth, with the exception of light. Radial growth decreased with increasing tree diameter, level of competition and proximity to skid trails, and varied among crown positions with trees in suppressed and intermediate positions having lower growth rates than codominants and dominants. Our results indicate that in selection managed stands, the radial growth of sugar maple trees depends on competition from close (≤6 m) conspecific neighbours, and is still affected by proximity to skid trails ∼10 years after harvesting. Such results underscore the importance of minimizing the extent of skid trail networks by careful pre-harvest planning of trail layout. We also conclude that the impact of heterogeneity among individual-tree neighbourhoods, such as those resulting from alternative spatial patterns of harvest, can usefully be integrated into models of post-harvest tree growth.     

参数平均方式对层状夹砂土壤积水入渗数值模拟的影响

为研究非饱和土壤水分运动方程中参数的平均方式对层状土水分运动数值模拟的影响,采用有限差分方法对Richards方程进行离散,并用MATLAB软件编程模拟层状夹砂土柱的薄层积水入渗过程。对差分网格中半节点参数的8种不同平均方式进行了数值模拟稳定性和精度的比较。结果表明:不同平均方式下湿润峰到达某一深度的时间与该处的土壤含水量增速呈负相关关系;间接算术平均和调和平均下的土壤含水率数值振荡较大;几何平均和三点平均2种方式下入渗率、累积入渗量和湿润锋误差相对较小。综合分析,几何平均和三点平均更适合层状夹砂土情况下薄层积水入渗过程的数值模拟。     

Improving estimates of population status and trend with superensemble models

Fishery managers must often reconcile conflicting estimates of population status and trend. Superensemble models, commonly used in climate and weather forecasting, may provide an effective solution. This approach uses predictions from multiple models as covariates in an additional “superensemble” model fitted to known data. We evaluated the potential for ensemble averages and superensemble models (ensemble methods) to improve estimates of population status and trend for fisheries. We fit four widely applicable data‐limited models that estimate stock biomass relative to equilibrium biomass at maximum sustainable yield (B/B_MSY). We combined these estimates of recent fishery status and trends in B/B_MSY with four ensemble methods: an ensemble average and three superensembles (a linear model, a random forest and a boosted regression tree). We trained our superensembles on 5,760 simulated stocks and tested them with cross‐validation and against a global database of 249 stock assessments. Ensemble methods substantially improved estimates of population status and trend. Random forest and boosted regression trees performed the best at estimating population status: inaccuracy (median absolute proportional error) decreased from 0.42 – 0.56 to 0.32 – 0.33, rank‐order correlation between predicted and true status improved from 0.02 – 0.32 to 0.44 – 0.48 and bias (median proportional error) declined from ?0.22 – 0.31 to ?0.12 – 0.03. We found similar improvements when predicting trend and when applying the simulation‐trained superensembles to catch data for global fish stocks. Superensembles can optimally leverage multiple model predictions; however, they must be tested, formed from a diverse set of accurate models and built on a data set representative of the populations to which they are applied.     

基于模型平均法的表层土壤含水率多模型综合反演

土壤水分是农作物和牧草生长的关键因素,也是影响全球气候变化和水循环的重要因子,因此准确监测土壤水分对促进农牧业可持续管理至关重要。基于Landsat 8 OLI多光谱影像,在表面生物物理特性和地形参数的基础上,结合野外实测土壤含水率,采用经验模型法分别构建了反距离加权法（Inverse Distance Weighting,IDW）、多元线性回归（Multivariable Linear Regression,MLR）、随机森林（Random Forest,RF）的土壤含水率反演模型,并采用模型平均法（Granger-Ramanathan,GR）进行组合反演,结果显示,通过模型平均法组合单一模型后在该研究区的土壤含水率反演中表现出了更好的适用性,模型平均法GR对于土壤含水率的估算效果佳,建模集与验证集R2≥0.88,均方根误差均不大于1.42%,且模型性能远优于反距离加权法、多元线性回归和随机森林。     

Effects of landscape-scale broadleaved woodland configuration and extent on roost location for six bat species across the UK

Although forest fragmentation can greatly affect biodiversity, responses to landscape-scale measures of woodland configuration in Europe have been examined for only a limited range of taxa. Almost all European bat species utilise woodland, however little is known about how they are affected by the spatial arrangement of woodland patches. Here we quantify landscape structure surrounding 1129 roosts of six bat species and a corresponding number of control locations across the UK, to examine associations between roost location and landscape composition, woodland proximity and the size of the nearest broadleaved woodland patch. Analyses are performed at two spatial scales: within 1 km of the roost and within a radius equivalent to the colony home-range (3–7 km). For four species, models at the 1 km scale were better able to predict roost occurrence than those at the home-range scale, although this difference was only significant for Pipistrellus pipistrellus. For all species roost location was positively associated with either the extent or proximity of broadleaved woodland, with the greatest effect of increasing woodland extent seen between 0% and 20% woodland cover. P. pipistrellus, Pipistrellus pygmaeus, Rhinolophus hipposideros, Eptesicus serotinus and Myotis nattereri all selected roosts closer to broadleaved woodland than expected by chance, with 90% of roosts located within 440 m of broadleaved woodland. Roost location was not affected by the size of the nearest broadleaved patch (patches ranged from 0.06–2798 ha ± 126 SD). These findings suggest that the bat species assessed here will benefit from the creation of an extensive network of woodland patches, including small patches, in landscapes with little existing woodland cover.