New features in the dendroTools R package: Bootstrapped and partial correlation coefficients for monthly and daily climate data

Climate-growth relationships are usually analysed using monthly climate data. The dendroTools R package also provides methodological approaches that enable climate-growth analysis for daily climate data. Such analysis reveals more complete climate signal patterns. In this article, new functions of the dendroTools R package are presented. Partial correlation coefficients are now implemented and can be used to calculate the strength of a linear relationship between two variables, while controlling for a third variable. Bootstrapped correlations can then be used to provide insights into the confidence intervals of statistical estimates. The calculation of partial and bootstrapped correlations is available for daily and monthly data. Finally, data transformation, S3 generic plotting and summary functions are also presented here.     

Comparison of an optimal regression method for climate reconstruction with the compare_methods() function from the dendroTools R package

The selection of a regression technique for climate reconstruction may have an important impact on reconstructed values. In this paper, we introduce the compare_methods() function from the dendroTools R package. This function compares different regression algorithms and returns validation results for each. In addition to mean validation metrics and ranks derived from these, transfer functions should have a key role in the evaluation of different regression algorithms. These are also returned as the output of compare_methods(). Our methodology is introduced on two case studies, one using a mean vessel area (MVA) chronology and one using a standardised tree-ring width (TRW) chronology. The nonlinear machine learning methods compared in our study provided relatively small (if any) improvements in terms of explaining climatic variance. However, they do offer different treatments of extreme values, and if providing more plausible climate reconstructions, this could make them a useful tool for climate reconstruction. We propose the use of the compare_methods() function as a standard methodological check before performing climate reconstruction.     

Applying climwin to dendrochronology: A breakthrough in the analyses of tree responses to environmental variability

Observational, correlative approaches are one of the backbones of dendrochronology. For instance, climate-growth relationships are usually quantified by calculating Pearson correlations. However, the ability to detect these relationships and the probability of declaring significant correlations by chance pose multiple challenges to such correlative framework. The R climwin package, developed a few years ago within the discipline of animal ecology, overcomes these limitations. In this paper we apply climwin to study relationships between climate and tree-ring widths and anatomy to show the advantages of using this package in the field of dendrochronology. This package allows calculating several models considering multiple windows relating a response variable to the climatic factors at different time resolutions. Then, the most parsimonious model is selected through an information-theoretic approach and randomization tests are computed to establish the significance of the selected model. We compare analyses based on Pearson correlations with climwin results using several environmental drivers (climate variables, drought indices, river flow), response variables (tree-ring width, tracheid lumen area and cell-wall thickness), and tree species from ecologically contrasting sites (cold- and water-limited conifers, Mediterranean riparian ash forests). Analyses of climate-growth/anatomy relationships based on the use of climwin showed several advantages over simple Pearson correlations: (i) they did not depend on the use of arbitrary time intervals of fixed duration, (ii) they allowed reducing probabilities associated with type I and II errors, (iii) they resulted in more consistent findings, (iv) they increased the capacity to detect differences between sites or periods in a time series, and (v) they provided more explanatory power.     

dendrometeR: Analyzing the pulse of trees in R

Dendrometers are measurement devices proven to be useful to analyze tree water relations and growth responses in relation to environmental variability. To analyze dendrometer data, two analytical methods prevail: (1) daily approaches that calculate or extract single values per day, and (2) stem-cycle approaches that separate high-resolution dendrometer records into distinct phases of contraction, expansion and stem-radius increment. Especially the stem-cycle approach requires complex algorithms to disentangle cyclic phases. Here, we present an R package, named dendrometeR, that facilitates the analysis of dendrometer data using both analytical methods. By making the package freely available, we make a first step towards comparable and reproducible methods to analyze dendrometer data. The package contains customizable functions to prepare, verify, process and plot dendrometer series, as well as functions that facilitate the analysis of dendrometer data (i.e. daily statistics or extracted phases) in relation to environmental data. The functionality of dendrometeR is illustrated in this note.     

川西亚高山不同海拔岷江冷杉树轮碳稳定同位素对气候的响应

树木年轮(简称树轮)碳稳定同位素技术是研究树轮气候学的一种有效方法。利用四川卧龙亚高山暗针叶林不同海拔高度岷江冷杉树轮样本资料,提取该树轮稳定碳同位素(δ13C)和去趋势序列(DS),研究其树轮碳稳定同位素序列对气候要素(降水、月平均温度和月平均相对湿度)的响应关系,初步揭示了在全球气候变化背景下,川西亚高山森林岷江冷杉树木生长对气候因子变化(气候变暖、降水减少等)的响应方式。主要结论有:(1)岷江冷杉树轮δ13C组成变化范围为-23.33‰—-26.31‰,平均值为-24.91‰,变异系数为-0.011—-0.038,并表现出较强的一阶自相关;其对环境变化有较好的指示作用,表明岷江冷杉树轮δ13C组成在年际变化中较为稳定。(2)低海拔的岷江冷杉树轮δ13C分馏主要与当年8月月平均相对湿度和当年12月月平均温度相关性显著(P0.05);高海拔岷江冷杉树轮δ13C分馏主要与上一年8月月平均相对湿度和当年4月月平均温度相关性显著(P0.05);中海拔的岷江冷杉树轮δ13C分馏主要与上一年1、11月月平均温度和当年2、11月月平均温度相关性显著(P0.05),冬季温度是中海拔区岷江冷杉树木生长的限制因子,且具有明显的"滞后效应"。川西卧龙亚高山暗针叶林岷江冷杉树木径向生长主要受到气温的制约,从生物学基础上阐明了树木生长与环境的关系,冬季温度的升高,有利于植物生长期的提前,植物生长旺盛,抗旱能力减弱;同时证明了建群种岷江冷杉对雨水的依赖很小,这有利于植物生存,且维持了该植物群落的稳定性。该研究弥补了我国青藏高原高海拔地区气象台站稀少、观测资料时间短缺,为预测未来气候变化对岷江冷杉树木径向生长变化提供了科学的依据。     

北亚热带马尾松年轮宽度与NDVI的关系

北亚热带地处暖温带向亚热带的过渡地区,对环境变化较为敏感。因此,研究北亚热带马尾松年轮宽度与森林NDVI的关系对于揭示陆地生态系统对全球气候变化的响应具有重要意义。以马尾松自然分布北界的南郑县和河南省鸡公山自然保护区为研究地点,利用北亚热带马尾松年轮宽度指数和1982-2006年逐月NOAA/AVHRR的归一化植被指数(NDVI)数据及气候数据,在分析年轮宽度及NDVI与气候因子关系的基础上,重点讨论了北亚热带马尾松径向生长与NDVI之间的关系。结果表明:北亚热带NDVI受水热条件的共同控制,其中与月均温相关性较强,且以正相关为主,与月降水量和干旱度指数多负相关;马尾松的径向生长与上一生长季的温度呈正相关,降水和干旱度指数为负相关,当年生长季内的温度和降水以促进作用为主,而与干旱度指数的关系在两地区内相反;南郑县和鸡公山地区年轮宽度与NDVI年值之间关系均不显著(P>0.05)。单月来讲,南郑县3、4、12月NDVI值与年表显著相关,鸡公山地区9月份的NDVI值与差值年表RES相关性最大;南郑县树木生长受温度影响最大,而鸡公山地区受温度和降水的综合作用。因此,在北亚热带地区,长时间序列的年轮宽度数据并不能很好反应NDVI的长期变化,利用树轮宽度指数来重建北亚热带地区NDVI需要进一步研究。     

Forward modeling analyses of Qilian Juniper <Emphasis Type="Italic">(Sabina przewalskii)</Emphasis> growth in response to climate factors in different regions of the Qilian Mountains,northwestern China

Key message

The process-based Vaganov–Shashkin model simulations accurately represent the nonlinear process of Qilian Juniper growth–climate relationship over different regions of the Qilian Mountains.

Abstract

The Qilian Mountains (QM), northeastern edge of the Tibetan Plateau (TP), is one of the most studied areas in China for tree-ring research; considerable dendroclimatic reconstructions have been carried out and a series of important achievements made over this region. However, most reconstructions were primarily based on empirical relationships between tree growth and climate factors identified through statistical analysis. Reliable information on the physiological processes of tree responses to climate change in different regions is still scarce. Here, the process-based Vaganov–Shashkin (VS) model was used to simulate regional patterns of climate–tree growth relationships using observed temperature and precipitation over the different regions of the QM. Results showed that the sequences of simulated growth curves were broadly consistent with the actual tree-ring chronologies in all three regions of the QM. VS model simulations accurately represented the effect of climatic controls on the growth of Qilian Juniper (Sabina przewalskii). VS model outputs closely corresponded to statistical relationships between tree-ring width and climate factors as well as observational physiological behavior; i.e., available water in May and June had the largest contribution to ring formation in Qilian Juniper over the QM. The simulated and actual data analyses revealed that the radial growth of trees in the western QM was more sensitive to moisture conditions in May and June, compared with growth in the central and eastern QM. Tree growth in this region is representative of large areas in northwestern China with drought conditions.

     

A comparison of metrics for estimating phylogenetic signal under alternative evolutionary models

Several metrics have been developed for estimating phylogenetic signal in comparative data. These may be important both in guiding future studies on correlated evolution and for inferring broad-scale evolutionary and ecological processes (e.g., phylogenetic niche conservatism). Notwithstanding, the validity of some of these metrics is under debate, especially after the development of more sophisticated model-based approaches that estimate departure from particular evolutionary models (i.e., Brownian motion). Here, two of these model-based metrics (Blomberg’s K-statistics and Pagel’s λ) are compared with three statistical approaches [Moran’s I autocorrelation coefficient, coefficients of determination from the autoregressive method (ARM), and phylogenetic eigenvector regression (PVR)]. Based on simulations of a trait evolving under Brownian motion for a phylogeny with 209 species, we showed that all metrics are strongly, although non-linearly, correlated to each other. Our analyses revealed that statistical approaches provide valid results and may be still particularly useful when detailed phylogenies are unavailable or when trait variation among species is difficult to describe by more standard Brownian or O-U evolutionary models.     

burnr: Fire history analysis and graphics in R

We developed a new software package, burnr, for fire history analysis and plotting in the R statistical programming environment. It was developed for tree-ring fire-scar analysis, but is broadly applicable to other event analyses (e.g., avalanches, frost rings, or culturally modified trees). Our new package can read, write, and manipulate standard tree-ring fire history FHX files, produce fire—demography charts, calculate fire frequency and seasonality statistics, and run superposed epoch analysis (SEA). A key benefit of burnr is that it enables automation of analyses and plotting, especially for large data sets. The package also facilitates creative plotting, mapping, and analyses when combined with the thousands of packages available in R. In this paper, we describe the basic functionality of burnr and introduce users to fire history analyses in R.     

Site-contingent responses to drought of core and relict Tetraclinis articulata populations from Morocco and Spain

The growth responses to climate variability are still unknown in locally threatened conifers from dry regions, but this information is necessary for improving the conservation of relict populations under increasing aridification. We characterized the radial growth patterns and responses to climate of Tetraclinis articulata, a Cupressaceae tree endemic to the western Mediterranean Basin, in a relict population located in southwestern Spain (Doñana) and two populations from the northern Morocco where the species core habitat is found (Tétouan, Ifarten). We assessed climate-growth relationships by using tree-ring width, climate data, drought and North Atlantic Oscillation (NAO) indices. Climate-growth analyses were refined using the climwin R package to select the most informative statistical models. The main climatic constraints of growth were inferred by using the process-based Vaganov-Shashkin (VS) model explicitly considering non-linear climate-growth relationships. Tetraclinis articulata growth was favored by wet conditions from the prior autumn to the spring of the growth year. In Doñana, warmer May conditions led to growth decline but this negative effect could be reversed by wet-warm conditions in the prior late autumn. Growth in the two Moroccan sites was constrained by 6- to 18-month long droughts peaking in summer, which account for cumulative water deficit since the previous autumn. Winter and early spring precipitation were the main climate drivers of growth in the Moroccan T. articulata populations, and their year-to-year variability was linked to the NAO. The VS model simulations showed that T. articulata growth is enhanced by wet soil conditions in late winter and early spring, probably recharging shallow soil water pools. The VS model also confirmed that warmer spring-summer conditions could amplify drought stress and threaten the long-term persistence of the relict Doñana population.     

Dendroclimatic calibration in R: The bootRes package for response and correlation function analysis

To understand the effect of climate on tree-ring features, such as width or density, tree-ring data have to be calibrated against instrumental records. The high degree of multicollinearity among monthly time series of climate data violates the assumption of independent predictor variables in ordinary least squares regression. Bootstrapped confidence intervals of parameter estimates via regression against the principle components of the predictor variables are a possible solution to that problem. Package bootRes for R implements a flexible interface for bootstrapped response and correlation function analysis and tackles some shortcomings of currently available software. Given the increasing popularity of the free R software for statistical analysis, bootRes should facilitate both using R as a computational environment among tree-ring scientists and implementing new approaches to dendroclimatic calibration.     

landscapemetrics: an open‐source R tool to calculate landscape metrics

Quantifying landscape characteristics and linking them to ecological processes is one of the central goals of landscape ecology. Landscape metrics are a widely used tool for the analysis of patch‐based, discrete land‐cover classes. Existing software to calculate landscape metrics has several constraints, such as being limited to a single platform, not being open‐source or involving a complicated integration into large workflows. We present landscapemetrics, an open‐source R package that overcomes many constraints of existing landscape metric software. The package includes an extensive collection of commonly used landscape metrics in a tidy workflow. To facilitate the integration into large workflows, landscapemetrics is based on a well‐established spatial framework in R. This allows pre‐processing of land‐cover maps or further statistical analysis without importing and exporting the data from and to different software environments. Additionally, the package provides many utility functions to visualize, extract, and sample landscape metrics. Lastly, we provide building‐blocks to motivate the development and integration of new metrics in the future. We demonstrate the usage and advantages of landscapemetrics by analysing the influence of different sampling schemes on the estimation of landscape metrics. In so doing, we demonstrate the many advantages of the package, especially its easy integration into large workflows. These new developments should help with the integration of landscape analysis in ecological research, given that ecologists are increasingly using R for the statistical analysis, modelling and visualization of spatial data.     

VS-oscilloscope: A new tool to parameterize tree radial growth based on climate conditions

It is generally assumed in dendroecological studies that annual tree-ring growth is adequately determined by a linear function of local or regional precipitation and temperature with a set of coefficients that are temporally invariant. However, various researchers have maintained that tree-ring records are the result of multivariate, often nonlinear biological and physical processes. To describe critical processes linking climate variables with tree-ring formation, the process-based tree-ring Vaganov–Shashkin model (VS-model) was successfully used. However, the VS-model is a complex tool requiring a considerable number of model parameters that should be re-estimated for each forest stand. Here we present a new visual approach of process-based tree-ring model parameterization (the so-called VS-oscilloscope) which allows the simulation of tree-ring growth and can be easily used by researchers and students. The VS-oscilloscope was tested on tree-ring data for two species (Larix gmeliniiand Picea obovata) growing in the permafrost zone of Central Siberia. The parameterization of the VS-model provided highly significant positive correlations (p < 0.0001) between simulated growth curves and original tree-ring chronologies for the period 1950–2009. The model outputs have shown differences in seasonal tree-ring growth between species that were well supported by the field observations. To better understand seasonal tree-ring growth and to verify the VS-model findings, a multi-year natural field study is needed, including seasonal observation of the thermo-hydrological regime of the soil, duration and rate of tracheid development, as well as measurements of their anatomical features.     

IncrementR: Analysing height growth of trees and shrubs in R

Dendrochronology mostly deals with secondary (radial) growth and attention to primary (height) growth has so far been limited. However, tree-ring widths might not adequately represent stem volume increments, net primary productivity and the size of the tree stem carbon sink. The main reason for the prevailing focus on radial growth is that establishing height growth chronologies requires time-consuming and destructive methods. However, for certain ecological applications, less laboriously acquired data on height growth averaged over several successive years are satisfactory. Here we present an R package that contains a set of tools for the analysis of height growth. The tools have been developed for input data of tree-ring widths extracted from series of successive stem height levels. Tree-ring widths ideally represent four directions in each cross section to capture potential changes in stem eccentricity between various height levels. The main computed parameters provided by the package include height growth along the stem, changes of stem eccentricity and stem taper. Accurate determination of average height growth depends on the correct estimation of the number of tree rings at different stem height levels, which might be complicated by missing rings in off-pith cores. The presented package therefore also contains functions implementing common procedures for the estimation of the number of missing tree rings near to the pith. Most outputs can be visualized graphically. The package is useful for estimating height growth in ecological and dendrogeomorphological studies, especially in situations where both primary and secondary growth is influenced by different environmental factors. It is also useful for analysing tree-ring chronologies assembled using serial sectioning, which typically applies to shrubs.     

Disentangling the roles of spatial and environmental variables in shaping benthic algal assemblages in rivers of central and northern China

Benthic algae were collected from central and northern Chinese rivers to test the hypothesis that geographic location has significant contributions in shaping algal assemblages. We used Moran’s eigenvector maps (MEM) to model spatial components and variation partitioning to quantify the influences of spatial and environmental variables on regional patterns of algal richness and community composition, respectively. We found that variation in algal richness was attributed to MEM component 2, 8, and 9 and the quadratic term of N–NO₃. Regarding abundance data, latitude, longitude, and MEM component 1, 2, and 7 were important spatial variables. Although P–PO₄, pH, and annual mean temperature were significant environmental variables influencing algal community composition, they were all spatially structured. Among the total explained variance in both algal metrics, spatial proportions were higher than that of environmental variables. We also found that abundant species of Achnanthidium minutissimum, Cocconeis placentula, Cymbella delicatula, Cymbella affinis, Cymbella turgidula, and Synedra ulna displayed clear spatially related patterns. In conclusion, the contributions of spatial and environmental variables to regional variation of algal assemblages are scale-dependent. As for our study scale (~1,000 km), spatial control may be more important. Since spatial effects could obscure local environmental impacts on algal communities, appropriate study scale and statistical methods should be taken into account in algal bioassessment. We recommend inclusion of both algal richness and community composition in study of algal biogeography, due to their different relationships with spatial and environmental variables.     

Individual and time-varying tree-ring growth to climate sensitivity of Pinus tabuliformis Carr. and Sabina przewalskii Kom. in the eastern Qilian Mountains, China

Individual tree-ring width chronologies and mean chronologies from Pinus tabuliformis Carr. (Chinese pine) and Sabina przewalskii Kom. (Qilian juniper) tree cores were collected and analyzed from two sites in the eastern Qilian Mountains of China. The chronologies were used to analyze individual and time-varying tree-ring growth to climate sensitivity with monthly mean air temperature and total precipitation data for the period 1958–2008. Climate–growth relationships were assessed with correlation functions and their stationarity and consistency over time were measured using moving correlation analysis. Individuals’ growth–climate correlations suggested increased percentages of individuals are correlated with certain variables (e.g., current June temperature at the P. tabuliformis site; previous June, December and current May temperature and May precipitation at the S. przewalskii site). These same climatic variables also correspond to the mean chronology correlations. A decreased percentage of individuals correlated with these climatic variables indicates a reduced sensitivity of the mean chronology. Moving correlation analysis indicated a significant change over time in the sensitivity of trees to climatic variability. Our results suggested: (1) that individual tree analysis might be a worthwhile tool to improve the quality and reliability of the climate signal from tree-ring series for dendroclimatology research; and (2) time-dependent fluctuations of climate growth relationships should be taken into account when assessing the quality and reliability of reconstructed climate signals.     

Identifying models of trait‐mediated community assembly using random forests and approximate Bayesian computation

Ecologists often use dispersion metrics and statistical hypothesis testing to infer processes of community formation such as environmental filtering, competitive exclusion, and neutral species assembly. These metrics have limited power in inferring assembly models because they rely on often‐violated assumptions. Here, we adapt a model of phenotypic similarity and repulsion to simulate the process of community assembly via environmental filtering and competitive exclusion, all while parameterizing the strength of the respective ecological processes. We then use random forests and approximate Bayesian computation to distinguish between these models given the simulated data. We find that our approach is more accurate than using dispersion metrics and accounts for uncertainty in model selection. We also demonstrate that the parameter determining the strength of the assembly processes can be accurately estimated. This approach is available in the R package CAMI; Community Assembly Model Inference. We demonstrate the effectiveness of CAMI using an example of plant communities living on lava flow islands.     

detrendeR – A Graphical User Interface to process and visualize tree-ring data using R

In this paper, we present the package detrendeR, a Graphical User Interface to facilitate the visualization and analysis of dendrochronological data, using the R computing environment. This package offers an easy way to perform most of the traditional tasks in dendrochronology: detrending, chronology building and graphical presentation of time series. The advantage of detrendeR, compared with the program ARSTAN, is the graphical interface that provides the user with an easy way to use R language, rich in graphics and handling routines, with no need to type commands. The detrendeR uses a simple and familiar dialog-box interface and it can read Tucson decadal-format files (*.rwl and *.crn) as well as plain text files. In addition, detrendeR has the ability to test temporal changes of the common signal using moving intervals. The detrendeR should make it easier to perform detrending and chronology building of tree-ring series, taking advantage of the R statistical programming environment.     

Stationarity of climate-growth response is only marginally influenced by the soil moisture regime in Western Siberia

Stationary (time-stable) relationships between a tree-ring proxy and climatic drivers are a prerequisite for using tree rings as paleo-climatological archives, but non-stationarity has been detected worldwide. Here we use a classical, temperature-sensitive treeline site in Western Siberia to specifically test the influence of micro-site conditions (wet versus dry) on the stationarity of climate-growth relationships in three co-existing conifer species: Larix sibirica Ledeb., Picea obovata Ledeb., and Pinus sibirica DuTour. We test two commonly used tree-ring proxies, annual tree-ring width (TRW) and maximum latewood density (MXD), using moving windows and the bootstrapped transfer function stability test (BTFS). Summer temperature is the main driver of tree growth in all three species, but non-stationarities exist in all species and both tree-ring proxies. For TRW, we found stationarity only for larch from both micro-sites, while for MXD, we found stationarity for spruce from both micro-sites, and for pine from the wet micro-site. Micro-site variability did not seem to affect stationarity in any systematic way. We highlight the necessity to systematically test the influence of different methods of stationarity tests, since BTFS was more sensitive than moving-window analysis. Taken together, our results underscore the importance of testing the assumption of stationarity for diverse micro-sites, different species and proxies at all sites prior to any tree-ring based temperature reconstruction, since even within one site results can be drastically different.     

Partly decoupled tree-ring width and leaf phenology response to 20th century temperature change in Sweden

The recent warming trend, and associated shifts in growing season length, challenge the principle of uniformitarianism, i.e., that current relations are persistent over time, and complicates the uncritical inferences of past climate from tree-ring data. Here we conduct a comparison between tree-ring width chronologies of Pinus sylvestris L. (Scots pine), Picea abies (L.) Karst. (Norway spruce) and Betula pubescens Ehrh. (Downy birch) and phenological observations (budburst and leaf senescence) of Fagus sylvatica L. (European beech), Quercus robur L. (European oak), Betula sp. (Birch), Norway spruce and Scots pine) in Sweden to assess to what extent the tree-ring width–temperature relationship and the timing of phenological phases are affected by increased temperature. Daily meteorological observations confirm a prolongation of the thermal growing season, most consistently observed as an earlier onset of around 1–2 weeks since the beginning of the 20th century. Observations of budburst closely mimic this pattern, with budburst of the deciduous trees occurring 1–2.5 weeks earlier. In contrast to the changes seen in phenology and observational temperature data, the tree-ring width–temperature relationships remain surprisingly stable throughout the 20th century. Norway spruce, Scots pine and Downy birch all show consistently significant correlations with at least one 30 day-long window of temperature starting in late June–early July season. Norway spruce displays the largest degree of stability, with a consistent 60 day-long temperature window with significant correlation starting around Julian calendar day 150. Thus, our results suggest that the principle of uniformitarianism is not violated during the period covered by modern meteorological observations. Further research is needed to determine at what thresholds the temperature sensitivity of these species may alter or deteriorate as a consequence of the ongoing climate change.