Methods for comparison of parameters from longitudinal rhythmometric models with multiple components

Multiple components linear least-squares methods have been proposed for the detection of periodic components in nonsinusoidal longitudinal time series. However, a proper test for comparison of parameters obtained from this method for two or more time series is not yet available. Accordingly, we propose two methods, one parametric and one nonparametric, to compare parameters from rhythmometric models with multiple components. The parametric method is based on techniques commonly and generally employed in linear regression analysis. The comparison of parameters among two or more time series is accomplished by the use of so-called dummy variables. The nonparametric method is based on bootstrap techniques. This approach basically tests if the difference in any given parameter obtained by fitting a model with the same periods to two different longitudinal time series differs from zero. This method calculates a confidence interval for the difference in the tested parameter. If this interval does not contain zero, it can be concluded that the parameters obtained from the two time series are different with high probability. An estimation of the p-value for the corresponding test can also be calculated. By the use of similar bootstrap techniques, confidence intervals can also be obtained for any parameter derived from the multiple component fit of several periods to nonsinusoidal longitudinal time series, including the orthophase (peak time), bathyphase (trough time), and global amplitude (difference between the maximum and the minimum) of the fitted model waveform. These methods represent a valuable tool for the comparison of rhythm parameters obtained by multiple component analysis, and they render this approach as a generally applicable one for waveform representation and detection of periodicities in nonsinusoidal, sparse, and noisy longitudinal time series sampled with either equidistant or unequidistant observations.     

Foliage litter turnover and earthworm populations in three beech forests of contrasting soil and vegetation types

Summary Leaf litter decomposition, levels of accumulated litter as well as the abundance and biomass of earthworms were measured in three mature beech forests in southern Sweden: one mor site, one poor mull site, and one rich mull site. The disappearance rate of beech litter, measured with litter bags, increased with increasing soil fertility. On the rich mull site, the disappearance rate was much higher than in the two other forests, due to the combined effects of higher earthworm activity, more favouable soil moisture conditions, and higher litter quality. Incubating the litter in finely meshed bags (1-mm mesh) to exclude macrofauna had a great effect on litter mass loss in the rich mull site, but it had only a minor effect in the other sites. Simultaneous incubations of local and transplanted leaf litter on the three study sites showed that the substrate quality of the litter increased in the order: mor site — poor mull site — rich mull site. Lignin, N, and P concentrations of the leaf litter failed to explain the observed differences in decomposition rates, and acid/base properties are suggested to be more important. Earthworm numbers per m² were 2.5 (1 species) in the mor, 40 (6 species) in the poor mull and 220 (9 species) in the rich mull forest. Soil chemical conditions, notably pH, were suggested as the main factors determining the inter-site differences in abundance and species composition of earthworms. The role of litter decomposition and earthworm activity in the accumulation of organic matter in the forest floor in different types of beech woodlands are discussed.     

Chemical composition of soil solutions extracted from New Zealand beech forests and West German beech and spruce forests

Concentrations of ions were measured in soil solutions from beech (Nothofagus) forests in remote areas of New Zealand and in solutions from beech (Fagus sylvatica) and Norway spruce (Picea abies) forests in North-East Bavaria, West Germany, to compare the chemistry of soil solutions which are unaffected by acid deposition (New Zealand) with those that are affected (West Germany). In New Zealand, soil solution SO₄ ^2– concentrations ranged between <2 and 58 mol L^–1, and NO₃ ^– concentrations ranged between <1 and 3 mol L^–1. In West Germany, SO₄ ^2– concentrations ranged between 80 and 700 mol L^–1, and NO₃ ^– concentrations at three of six sites ranged between 39 and 3750 mol L^–1, but was not detected at the remaining three sites. At all sites in New Zealand, and at sites where the soil base status was moderately high in West Germany, pH levels increased, and total Al (Al_t) and inorganic monomeric Al (Al_i) levels decreased rapidly with increasing soil depth. In contrast, at sites on soils of low base status in West Germany, pH levels increased only slightly, and Al levels did not decline with increasing soil depth.Under a high-elevation Norway spruce stand showing severe Mg deficiency and dieback symptoms in West Germany, soil solution Mg²⁺ levels ranged between 20 and 60 mol L^–, and were only half those under a healthy stand. Al_t and Al_i levels were substantially higher the healthy stand than under the unhealthy stand, indicating that Al toxicity was not the main cause of spruce decline.     

Imputation and variable selection in linear regression models with missing covariates

Across multiply imputed data sets, variable selection methods such as stepwise regression and other criterion-based strategies that include or exclude particular variables typically result in models with different selected predictors, thus presenting a problem for combining the results from separate complete-data analyses. Here, drawing on a Bayesian framework, we propose two alternative strategies to address the problem of choosing among linear regression models when there are missing covariates. One approach, which we call "impute, then select" (ITS) involves initially performing multiple imputation and then applying Bayesian variable selection to the multiply imputed data sets. A second strategy is to conduct Bayesian variable selection and missing data imputation simultaneously within one Gibbs sampling process, which we call "simultaneously impute and select" (SIAS). The methods are implemented and evaluated using the Bayesian procedure known as stochastic search variable selection for multivariate normal data sets, but both strategies offer general frameworks within which different Bayesian variable selection algorithms could be used for other types of data sets. A study of mental health services utilization among children in foster care programs is used to illustrate the techniques. Simulation studies show that both ITS and SIAS outperform complete-case analysis with stepwise variable selection and that SIAS slightly outperforms ITS.     

中国东北地带性顶极植被类型及其预测判别模型——动态地植物学说的继承与发展(Ⅰ)

本文根据刘慎谔的动态地植物学说原理,采用主成分分析法(PCA)和多组判别分析法(MGDA)研究了东北地区地带性顶极植被类型与气候指标的关系:1)以东北地区210个气象站的7个气象因子为变量,组成210×7矩阵,用PCA方法进行分类和排序,其结果明显划分7个地带性顶极植被类型,2)影响地带性顶极植被类型的形成和分布的气候因子,主要是热量和水分条件以及二者的组合状况;3)确定了7个地带性顶极植被类型的水热指数分布范围和特征;4)采用MGDA法建立了判别函数模型,预测了东北地区地带性顶极植被类型,其准确率可达90.32％。     

不同菠萝品种种植对连作蕉园土壤理化性质和可培养微生物数量的影响

【背景】不同作物轮作是克服作物连作障碍的重要措施,香蕉与菠萝轮作能有效缓解香蕉土传枯萎病。【目的】以休耕(CK)、高病蕉园土壤中种植巴西香蕉苗(B)、高病蕉园土壤中种植"巴厘"菠萝苗(B_BP)、高病蕉园土壤中种植"金菠萝"菠萝苗(B_GP)和高病蕉园土壤中种植"台农17号"菠萝苗(B_PP)为对象,研究不同菠萝品种种植对连作蕉园土壤理化性质和可培养尖孢镰刀菌、真菌、细菌和放线菌数量的影响,揭示不同菠萝品种种植在高病蕉园土壤中理化性质差异及微生物分布特征。【方法】采用盆栽试验结合可培养微生物研究方法,探究在高病蕉园土壤种植不同菠萝品种后的土壤理化性质和可培养微生物数量的变化。【结果】与休耕(CK)处理相比,高病蕉园土壤中继续种植巴西香蕉苗(B)处理显著增加土壤中尖孢镰刀菌数量,而种植菠萝品种"金菠萝"和"台农17号"(B_GP和B_PP)处理均能显著降低土壤中可培养尖孢镰刀菌数量,增加细菌和放线菌数量。土壤速效磷、细菌及放线菌数量均与可培养尖孢镰刀菌数量呈显著负相关关系;pH和真菌数量与可培养尖孢镰刀菌数量呈显著正相关关系。主坐标分析(principalco-ordinatesanalysis,PCoA)和多元回归树(multivariate regression trees,MRT)分析结果表明,种植"金菠萝"和"台农17号"菠萝(B_GP和B_PP)处理间的土壤肥力质量相近,并显著区别于其他3个处理。【结论】在高发病香蕉园地,种植菠萝品种"金菠萝"和"台农17号"(B_GP和B_PP)可以显著改善其土壤理化性质和土壤可培养微生物状况,对香蕉连作障碍有较好的缓解作用。     

Bivariate versus multivariate allometry: A note on a paper by Jungers and German

Jungers and German (1981) found differences when they compared 1) coefficients of allometry from bivariate plots of log measurements versus log body weight with 2) those coefficients from the first principal component of the log measurements excluding body weight. It is argued here that an arbitrary choice of unit for “internal size” is all that separates these coefficients. When the unit is chosen to make internal size isometric with body weight the coefficients agree rather well.     

Correlations among vegetation strata and environmental variables in subalpine spruce-fir forests,southeastern British Columbia

A set of forest vegetation-environment data consisting of 110 plots from subalpine spruce-fir forests in southeastern British Columbia was analyzed with multivariate methods. The aims were to describe and compare relationships among plots using species composition data from four strata of the forest (trees, shrubs, herbs, bryophytes), and to compare patterns in the separate strata with patterns in the environmental data. Weak but statistically significant correlations were detected among PCA axes from all four strata, with a tendency for correlations to be stronger between axes of adjacent strata. Relationships between stratal patterns and environmental variation were examined by correlating PCA axes with environmental variables separately, and jointly with canonical correlation analysis. Linear composites of the environmental data identified by the first canonical axis were similar for shrub, herb, and bryophyte strata implying similar responses to environmental variation in understorey strata; however, < 16% of the total variance was explained. No clear relationships emerged between tree stratum and environmental variation. Probable determinants of pattern in the tree stratum include historical events such as fire and winter snowpack, as well as possible interactions between species. Collectively, these events bear little resemblance to existing environmental conditions in the forest.Nomenclature used by BCMF follows Crum, Steere & Anderson (1973), Hale & Culberson (1970), and Stottler & Crandall-Stottler (1977) for non-vascular plants, and Taylor & MacBryde (1977) for vascular plants.The British Columbia Ministry of Forests provided the data on which this paper is based, as well as a grant to pay for the computer time used. We are grateful to G. Utzig for initially suggesting that a project of this nature be done, and to S. Phelps for helping with the retrieval of computer data files from tape. This paper has greatly benefited from discussions with P. Comeau and R. K. Scagel, and the comments of two anonymous reviewers.     

Variation of soil and biomass carbon pools in beech forests across a precipitation gradient

Temperate forests have recently been identified as being continuing sinks for carbon even in their mature and senescent stages. However, modeling exercises indicate that a warmer and drier climate as predicted for parts of Central Europe may substantially alter the source/sink function of these economically important ecosystems. In a transect study with 14 mature European beech (Fagus sylvatica L.) forests growing on uniform geological substrate, we analyzed the influence of a large reduction of annual precipitation (970–520 mm yr^?1) on the carbon stocks in fast and slow pools, independent of the well‐known aging effect. We investigated the C storage in the organic L, F, H layers, the mineral soil to 100 cm, and in the biomass (stem, leaves, fine roots), and analyzed the dependence of these pools on precipitation. Soil organic carbon decreased by about 25% from stands with > 900 mm yr^?1 to those with < 600 mm yr^?1; while the carbon storage in beech stems slightly increased. Reduced precipitation affected the biomass C pool in particular in the fine root fraction but much less in the leaf biomass and stem fractions. Fine root turnover increased with a precipitation reduction, even though stand fine root biomass and SOC in the organic L, F, and H layers decreased. According to regression analyses, the C storage in the organic layers was mainly controlled by the size of the fine root C pool suggesting an important role of fine root turnover for the C transfer from tree biomass to the SOC pool. We conclude that the long‐term consequence of a substantial precipitation decrease would be a reduction of the mineral soil and organic layer SOC pools, mainly due to higher decomposition rates. This could turn temperate beech forests into significant carbon sources instead of sinks under global warming.     

Prediction of protein solubility in Escherichia coli using logistic regression

In this article we present a new and more accurate model for the prediction of the solubility of proteins overexpressed in the bacterium Escherichia coli. The model uses the statistical technique of logistic regression. To build this model, 32 parameters that could potentially correlate well with solubility were used. In addition, the protein database was expanded compared to those used previously. We tested several different implementations of logistic regression with varied results. The best implementation, which is the one we report, exhibits excellent overall prediction accuracies: 94% for the model and 87% by cross‐validation. For comparison, we also tested discriminant analysis using the same parameters, and we obtained a less accurate prediction (69% cross‐validation accuracy for the stepwise forward plus interactions model). Biotechnol. Bioeng. 2010; 105: 374–383. © 2009 Wiley Periodicals, Inc.     

Prediction of occurrence of vascular plants in deciduous forests of South Sweden by means of Ellenberg indicator values

Abstract. In this study we present a new method for predicting the occurrences of species using data from deciduous forests in South Sweden. Complete species lists of vascular plants were compiled from 101 stands and from representative sample plots inside the stands. Soil samples from each stand were collected for determination of pH and nitrogen mineralization. Presence-absence data for species were fitted to the values of four environmental variables - soil moisture, soil reaction (pH), soil nitrogen and light - by means of Linear (Multiple) Logistic Regression (LLR), and Gaussian (Multiple) Logistic Regression (GLR). First, these values were estimated by calculating the weighted averages of Ellenberg indicator values. Second, the estimates for reaction and nitrogen were substituted by the real measurements of pH and mineralized NH₄⁺, keeping the Ellenberg estimates for light and moisture. The models were validated by an independent test data set. In general, the models had high predictive abilities. GLR fitted the species occurrences better to the environmental variables than LLR, but had a lower accuracy of prediction of species occurrence in the stands. The use of soil measurements instead of Ellenberg indicator values did not improve the predictive abilities of the models. The environmental conditions in the stand test set were successfully estimated by using species data from the plots. When using the species lists of the stands instead of plot data, a slightly better predictive ability was obtained. The collection of plot data, however, is easier and less time-consuming. The accuracy of prediction differed considerably between species.     

Relationships between the vegetation and soil conditions in beech and beech-oak forests of northern Germany

In this study we examine the relationships between the vegetation of beech and beech-oak forest communities (Hordelymo-Fagetum, Galio-Fagetum, Deschampsio-Fagetum, Betulo-Quercetum) and their soil conditions in the lowlands of northern Germany, based on 84 sample plots. In all plots the vegetation was recorded and soil parameters were analysed (thickness of the O- and the A-horizons, pH, S-value, base saturation, C/N, mean Ellenberg moisture indicator value). The vegetation classification according to the traditional Braun-Blanquet approach was compared with the result of a multivariate cluster analysis. Vegetation-site relationships were analysed by means of an indirect gradient analysis (DCA).Both traditional classification methods and the cluster analysis have produced comparable classification results. So far as the species composition is concerned, a similar grouping of sample plots was found in both approaches. Multivariate cluster analysis thus supports the classification found by the Braun-Blanquet method. The result of the DCA shows that the four forest communities mentioned above represent clearly definable ecological units. The main site factor influencing changes in the species composition is a base gradient, which is best expressed by the S-value. In addition, within the series Hordelymo-Fagetum - Galio-Fagetum - Deschampsio-Fagetum the C/N-ratios and the thickness of the organic layers (O-horizon) increase continuously. By contrast, the floristic differences between oligotrophic forest communities (i.e., Deschampsio-Fagetum and Betulo-Quercetum) cannot be explained by a base gradient and increasing C/N-ratios. It is suggested that a different forest management history in some cases (e.g., promotion of Quercus robur by silvicultural treatments) is responsible for differences in the species composition, but on the other hand the result of the DCA indicates that Fagus sylvatica is replaced by Quercus robur with increasing soil moisture (i.e., with the increasing influence of a high groundwater table). Summarizing these results, it can be concluded that the ecological importance of single site factors affecting the species composition changes within the entire site spectrum covered by the beech and beech-oak forests of northern Germany.     

Path analysis of standing crop and environmental variables in the field layer of two Belgian riverine forests

Seasonal variation in standing crop, litter, pH and relative light intensity was measured in the field layer of two riverine forests near Bruges, Belgium. At each site the living fraction in 10 randomly selected 0.25 m² quadrats was separated into various components and graphed in relation to the environmental variables. The effect of the environmental variables on standing crop was assessed using path analysis, based on the standardized partial regression coefficients obtained in separate stepwise multiple regressions. In order to linearize the relation between the standing crop and the environmental variables, and to meet other statistical assumptions, a log-transformation was applied to the standing crop data. On an annual basis, 44% of the total variation in log-standing crop was explained by the variables litter, relative light intensity and pH, while seasonally the coefficient of determination ranged from 0.23 (November) to 0.72 (April). A general path diagram, based on all data, indicated predominant direct effect of litter (path coefficient: ?0.55) and pH (0.26). Relative light intensity had only an indirect effect through its impact on litter and pH. Considerable seasonal changes occurred in the importance of litter, relative light intensity and pH.     

Fish ecoregions of Kansas: stream fish assemblage patterns and associated environmental correlates

Synopsis Principal components analysis was performed on fish presence/absence data for 39 common fish species from 410 stream sites in Kansas. The analysis confirmed ten ecologically meaningful fish assemblages, based on species associations. Factor scores based on these assemblages were then clustered into six geographic areas or fish ecoregions. Canonical discriminant analysis identified environmental variables that distinguished the derived fish ecoregions. Mean annual runoff, mean annual growing season, and discharge appear most important. Mean width, mean depth, chloride concentration, water temperature, substrate type, gradient, and percent of pool habitat were less important. Correspondence exists between these fish ecoregions and the patterns of physiographic regions, river basins, geology, soil, and potential natural vegetation in Kansas. The multivariate statistical approach used to classify fish ecoregions should have considerable potential value for fish assessment and management purposes in areas other than the state of Kansas.     

Dynamic floristic changes of Swedish beech forest in relation to soil acidity and stand management

This study reports dynamic changes in the beech forest vegetation during one decade, using 95 permanent observation areas representing a wide variety of soils and management regimes. Current soil acidification, including decreasing pH and base cation pools, increasing solubility of toxic elements and increasing deposition of N, as well as recent changes in the beech forest management have created good conditions for the study. Most species of vascular plants increased their frequencies during the 1980's. However, there were several notable exceptions, in particularGalium odoratum, Viola riviniana/reichenbachiana, Polygonatum multiflorum, andMercurialis perennis. These species, demanding a comparatively low soil acidity for survival, are distinctly disfavoured by the long-term soil changes in the forests, which seem to have approached or exceeded their limits of existence in many sites. With most other species, differences in management regimes between the beginning and the end of the observation period were more important to the frequency changes. Sensitive to heavy thinning of the stands were, e.g.Oxalis acetosella, Lamium galeobdolon andMelica uniflora, favoured by thinning wereStellaria nemorum, Carex pilulifera, Milium effusum and an appreciable number of more ephemeral species normally occurring in clear-cut areas or otherwise open land, e.g.,Rubus idaeus, Galeopsis tetrahit, Athyrium filix-femina, Juncus effusus, Agrostic capillaris, Veronica officinalis, Urtica dioica, andMoehringia trinervia. Saplings of woody plants usually also became more frequent during the 1980's.     

Decomposition of <Superscript>15</Superscript>N-labelled beech litter and fate of nitrogen derived from litter in a beech forest

The decomposition and the fate of ¹⁵N- labelled beech litter was monitored in a beech forest (Vosges mountains, France) over 3 years. Circular plots around beech trees were isolated from neighbouring tree roots by soil trenching. After removal of the litter layer, ¹⁵N-labelled litter was distributed on the soil. Samples [labelled litter, soil (0–15 cm depths], fine roots, mycorrhizal root tips, leaves) were collected during the subsequent vegetation periods and analysed for total N and ¹⁵N concentration. Mass loss of the ¹⁵N-labelled litter was estimated using mass loss data from a litterbag experiment set up at the field site. An initial and rapid release of soluble N from the decomposing litter was balanced by the incorporation of exogenous N into the litter. Fungal N accounted for approximately 35% of the N incorporation. Over 2 years, litter N was continuously released and rates of N and mass loss were equivalent, while litter N was preferentially lost during the 3rd year. Released ¹⁵N accumulated essentially at the soil surface. ¹⁵N from the decomposing litter was rapidly (i.e. in 6 months) detected in roots and beech leaves and its level increased regularly and linearly over the course of the labelling experiment. After 3 years, about 2% of the original litter N had accumulated in the trees. ¹⁵N budgets indicated that soluble N was the main source for soil microbial biomass. Nitrogen accumulated in storage compounds was the main source of leaf N, while soil organic N was the main source of mycorrhizal N. Use of ¹⁵N-labelled beech litter as decomposing substrate allowed assessment of the fate of litter N in the soil and tree N pools in a beech forest on different time scales. Received: 3 May 1999 / Accepted: 3 January 2000     

红松混交林凋落物氮储量及分解释放对土壤氮的影响

2012年5—10月,采用直接收获法,研究了小兴安岭地区云冷杉红松混交林和枫桦红松混交林两种林型凋落物的未分解层(L层)、半分解层(F层)和腐殖质层(H层)以及土壤表层(S层)氮储量及凋落物分解释放对土壤氮影响。结果表明:研究期间两种林型凋落物现存量变化范围分别为19.43~27.25和21.25~24.28 t·hm-2,氮储量变化范围分别为287.21~418.22和274.81~351.21 kg·hm-2,各层氮含量大小次序均为LFHS;云冷杉红松混交林各层凋落物现存量及其氮储存量5月和9月达到峰值,每月氮储量从L~H层均增加,凋落物分解释放氮在F和H层易富集,输入到土壤中较少;枫桦红松混交林各层凋落物现存量及其氮储量5月和10月达到峰值,每月氮储量从L~H层均减少,氮在凋落物各层中均易迁移,输入到土壤中的氮比云冷杉红松混交林多;两种林型L、F、H层凋落物现存量以及H层氮含量与S层氮含量之间,L和F层凋落物现存量与H层氮含量之间均呈显著正相关。