Forecasting community dynamics: the mast system

Forecasting community dynamics: the mast system

Overview

PIs: Jim Clark, Jen Swenson, Roland Kays

The impact of climate change on biological communities will depend on interactions involving the local habitat and the species that interact with one another—as each species responds directly to climate it indirectly affects all of the species with which it interacts. These species interactions complicate our ability to predict climate effects, because each species experiences habitat complexity at a different scale—from flightless insects to large vertebrates.  Current efforts focus on the effects of climate change, land cover, and soils, but do not benefit from estimates of food availability. This study will determine how diverse communities of species monitored in NEON (ground beetles, vascular plants, small mammals, birds) respond together with food supply, in the form of masting shrubs and trees, and large mammal surveys. A focus on the mast system of pulsed seed and fruit production from trees, includes vertebrate consumers, and indirect interactions with arthropod competitors and vertebrate predators. Remotely-sensed imagery and the NEON airborne observatory will be used to characterize habitat diversity.

Results of this analysis will be used to evaluate community change and reorganization, including prediction and attribution of climate risk by species and habitat and how it is shared across species groups. New data on large mammals and seed production from NEON sites will be made available to the community. The study will engage the public through citizen assisted identification of animal images.

View from a Yellowstone plot

Field Notes from NEON sites

Our study aims to evaluate the contribution of mast (fruit and seed production by trees and shrubs) to consumer abundances, relative to other food sources.  Specifically, we deploy seed traps, evaluate individual tree attributes (include cone production for conifers), and camera traps, to quantify activity of large vertebrates. This collaboration includes Jen Swenson at Duke University and Roland Kays at NCSU and the Museum of Natural History.

At the time of this project, NEON sites are just beginning to host visits from individual PIs.  Each site has its own ownership, permitting system, and management plan. For example, Florida sites include active fire management, which affects any sampling equipment left in the field.  Permits and sampling are negotiated with each site individually.  The time required to obtain permits varies site-to-site.  These metadata notes summarize some of these issues for sites we sampled to be used by us, but perhaps by others.

The Disney Wilderness Preserve (DSNY)

15-16 June 2018

Jim Clark, Jordan Luongo

Sampling at the DSNY NEON site was facilitated by Guy Fausnaught at NEON and Beatriz Pace-Aldana, the Research Coordinator, plant and GIS specialist here.  From 15-16 June we installed seed traps and collected data on cone production at three NEON plots: 6, 8, 10.  Tree cover on other plots was too sparse to warrant seed production.

 

Formerly ranchland, the Disney Wilderness Preserve has been restored to Pine savanna under ownership of The Nature Conservancy.  Prescribed burning at 3-yr intervals has transformed pastures to pine savannas, interspersed with cypress swamps and some hardwood stands.  TNC is actively managing to minimize invasive species and restore red-cockaded woodpecker to the site.  The dominant slash pine (Pinus elliotii) and longleaf (P. palustris) form a sparse canopy over palmetto (Serenoa repens) and rhizomatous dwarf shrubs, including dwarf live oak (Quercus minima), runner oak (Q. elliottii), and blueberry (Vaccinium) and grass species, include wiregrass (Aristida stricta).  We observed a single loblolly pine (P. taeda).  Much of the mesic flatwoods remain saturated from May to September from almost daily afternoon thunderstorms, sometimes short-duration, often high-intensity.

Pine savannas at Disney Preserve

On arrival, lab Manager Jordan Luongo and I were briefed by Beatriz on site history, layout, and protocols.  She discussed options for accommodating our seed traps during controlled burns, which can come with limited notice.  She generously offered to facilitate moving our traps aside during burns and contacting us on how to redeploy them.  Our reaction to individual burns will be complicated by the distance and travel expense.

 

As is typical for this time of year, many of the roads held up to ½ m of water; 4WD is essential.  We worked around the typical thunderstorms, establishing seed traps at three NEON sites 6, 8, and 10.

 

Cone counts supplement mast estimates from seed trap data, using the model and code in the R package MASTIF.  Throughout this site, pines support sparse foliage and cone production.  Female cones fertilized in spring develop over the current and subsequent growing season to release seeds in autumn and winter of the following year. Sparse canopies facilitated direct cone counts, in three cohorts.  Open cones at the time of our visit in 2018 released seeds in winter of 2017.  Consistent with previous convention, we call this the “2017 seed year”.  Fully developed but unopened cones on trees at the time of our visit were initiated in spring 2017 and will release seed beginning in autumn of 2018.  Undeveloped seeds were initiated in 2018.  From cone counts, it became clear that 2017 was a relatively strong seed year. From sparse unopened cones we expect small seed crops in 2018 and 2019.

 

Wildlife observed by us included white-tailed deer, diverse songbirds, Florida sandhill cranes, great blue heron, little blue heron, cattle egret, red-shouldered hawk.  Evidence of feral pig activity is common.

 

Plot 10 represents palmetto-dominated savannas, with several shrub oaks, Vaccinium, and grasses.  The water table was at the surface, but there was no standing water within the plot.  We installed 6 seed traps and counted cones from the three cohorts on trees within 20 m of the plot center.

 

Having been burned the week before our arrival, Plot 6 presented us with blackened soils and stems. Already there was evidence of regrowth of palmetto and grasses.  Despite near complete consumption of understory foliage during the burn, it was clear that palmetto was a dominant species. We installed six seed traps and counted cones on trees.

 

At the southern end of the site, plot 8 supported a diverse understory beneath a sparse slash pine overstory.  Here again, we completed six seed traps and cone counts.

 

 

Ecological diversity and climate change (ENV 623L, Fall 2018)

Ecological diversity and climate change (ENV 623L, Fall 2018)

Anticipating climate effects on biodiversity challenges ecologists to translate information on individual organisms and processes to ecological dynamics. How are ecologists putting the information together? How much is supported by mechanistic evidence? By theory? Are experiments and observational studies contributing in different ways? The tools applied to interpret climate effects span the foundations in jointly distributed random variables to simulation and machine learning. The two goals of this course address the confluence of science and emerging tools to interpret the evidence.

The first goal is to evaluate the science of biodiversity and climate change, including the changes that are happening now, in the past, and what we can expect in the future.

Secondarily, the course provides an overview of analytical tools. However, we reverse the traditional approach–rather than describe a model, then look for an application, we start with compelling controversies involving global change impacts.

The format combines lecture material, weekly labs, literature discussion, and individual and group assignments.

BiodiverClimClass

Topics from 2017:

Wrangling big data from the internet

Trophic cascades in Yellowstone National Park

The impact of fishing technology on long-term trends in fish returns

Climate variation and its effects on phenology

Climate versus habitat effects on biodiversity in the NEON network

Climate versus soil effects on tree species distribution and abundance in the NEON network

 

mastif

mastif

mast inference and forecasting (mastif)

mastif is used to estimate fecundity and seed dispersal based on mapped tree and seed trap data.  mastif operates on inventory plot data or simulated data.  Inventory data for seeds of the genus Pinus are shown below for random plots and years.  In these maps, large green symbols indicate seed traps that accumulated large numbers of seeds.  Large brown symbols indicate large trees.

The simulator in mastif produces data sets with known parameter values and latent variables.  Analysis provides insight on identifiability of seed production and the variables that control it.

At left is an example of seed prediction and fecundity estimates with known values from simulation.  Simulated data are generated with the function mastSim.

 

 

 

 

Prediction validates the model and shows how well parameters and latent states are identified for specific data sets.

 

 

 

 

Maps of predicted seed rain and uncertainty from the fitted model.  Predicted mean seed rain at left and predictive coefficient of variation at right.  In-sample prediction includes mapped stands fitted with the model.  Out-of-sample prediction is available for mapped stands that are not fitted to the data.

 

 

Predicted mean seed rain, showing seed traps with symbols in proportion to seed counts.
Predictive coefficient of variation for seed rain, showing highest uncertainty in regions with few seed traps.

 

 

 

 

 

Interannual variability in fecundity at the individual tree scale differs by species and regions, showing pines at Duke Forest (DF), Coweeta (CW), and Harvard Forest (HF). In the plot at left are year effects for several species and regions, shown as 95% credible intervals with seed count data (lower dots with mean values).

 

 

 

 

 

 

 

 

 

Year effects show ‘masting’, the tendency for individuals to synchronize reproduction.  Year effects can be fitted by species and by species and region.

 

 

 

 

 

mastifuses Gibbs sampling with direct sampling, Metropolis, and Hamiltonian updating.  It is implemented in R and C++ with Rcpp and the RcppArmadillo library for fast, efficient linear algebra operations. For this development version you will need to install them. Occasionally this is challenging on some platforms–if you have trouble send me an email. Someone in the lab has probably figured it out for a machine like yours. This will not be an issue once it is on CRAN.

 

Vignette with R code and applications

Package available here: mastif_1.0.tar.gz

Installation in RStudio:

> install.packages('mastif_1.0.tar.gz',repos=NULL, type='source')
> library('mastif')

Documentation:

> help('mastif')
> browseVignettes('mastif')
R packages

R packages

gjam – generalized joint attribute modeling

mastif – mast inference and forecasting

schedule and readings

schedule and readings


week unit example concepts applications
1 (1/12) introductory concepts observations/experiments/models; forward and backward: inference/prediction/simulation; terms like frequentist/Bayesian/algorithmic; false dichotomy of process/statistical basic notation; try R; USDA Forest Inventory and Analysis Program
2 (1/17) big data R basics applied to big data; strategies for hundreds of files to data exploration to analysis; data structures/formats; types of variables; data exploration Breeding bird survey
2 (1/24) joint distributions conditional, marginal, joint; Bayes theorem; total probability; conditional independence; an observation as a random variable, a random sample as a joint distribution experimental pathogen co-infection
3 (1/31) model taxonomy, construction functions, parameters, types of error; types of variables; discrete vs stochastic; discretizing continuous variables; ANOVA as a model for design NEON biol data
4 (2/7) elements of inference likelihood and maximum likelihood; Fisher information, confidence intervals; prior and posterior distributions; predictive distributions; conjugacy; variable selection TBA
5 (2/14) linear models, GLMs design matrix, rank, role in estimates/information; covariates and factors; main effects, interactions; standardization; least squares as ML; Bayesian analysis of LM; distribution of data; random effects TBA
6 (2/21) (Jim gone) GLMs logit, probit; Poisson regression; Ordinal regression; applications TBA
7 (2/28) hierarchical analysis the Bayesian paradigm; hierarchical structures; revisit random effects; applications: mixed model regression atmospheric, ocean circulation
8 (3/7) more on computation MCMC; Gibbs sampling; Metropolis; posterior and predictive distributions TBA
9 (3/14) (spring break)
10 (3/28) multivariate responses co-dependence in MVN, multinomial; mean vs covariance structure; count data in the multinomial or a 1st-stage Poisson; GJAM application to NEON breeding bird survey, traits, NEON
11 (4/4) space and time state-space models; spatial random effects phenology, invasion, disease, LiDAR
12 (4/11) presentations

Readings

1 Bayesian modeling and computation

Why Big Data Could Be a Big Fail, Jordan on potential and limitations of Big Data (confusing title).

Why environmental scientists are becoming Bayesians, Clark on expanding use of Bayes, Ecol Letters.

Bayesian method for hierarchical models: Are ecologists making a Faustian bargain? Lele and Dennis offer contrarian view, Ecol Appl.

2. big data, try R

Tidy Data, Wickam on concepts that link models and data to algorithms to R, J Stat Soft.

Connectivity of wood thrush breeding, wintering, and migration sites, Stanley et al on tracking wood thrush populations combined with BBS Cons Biol.

3. joint distributions

Models for Ecological Data, Appendix D.

Evaluating the impacts of fungal seedling pathogens on temperate forest seedling survival, Hersh et al. on joint, conditional, predictive distributions, Ecology.

4. model taxonomy/construction

Models for Ecological Data, chapter 3.

Accounting for uncertainty in ecological analysis: the strengths and limitations of hierarchical statistical modeling, Cressie et al. review modeling issues and challenges

Combining statistical inference and decisions in ecology., Williams and Hooten on assembling and using models Ecol Appl.

5. elements of inference

Models for Ecological Data, chapter 4.

readings for previous unit

6. linear models, GLMs

Models for Ecological Data, chapter 5, 6.

Uncertainty Management in Coupled Physical-Biological Lower Trophic Level Ocean Ecosystem Models, Gelman and Park on regression with context of group effects, the ‘ecological fallacy’, J Roy Stat Assoc.

7. GLMs

Models for Ecological Data, chapter 7.

8. hierarchical analysis

Models for Ecological Data, chapter 8.

Assessing abiotic conditions influencing the longitudinal distribution of exotic brown trout, Meredith et al. with ZINB to describe stream occupancy, Biol Invasions.

Hierarchical models of animal abundance and occurrence, Royle and Dorazio. J Agric, Biol, Environm Stat.

Uncertainty Management in Coupled Physical-Biological Lower Trophic Level Ocean Ecosystem Models, Wikle et al. describe the connections between a physical model and data, Oceanography.

9. more on computation

Models for Ecological Data, chapter 7, 8.

10. multivariate responses

Species distributions

Stacking species distribution models and adjusting bias by linking them to macroecological models, Calabrese et al. on modeling multiple species distributions, Global Ecol Biogeogr.

Hierarchical species distribution models. Hefley and M.B. Hooten review models, Current Landscape Ecol Rep.

Estimating the Effects of Habitat and Biological Interactions in an Avian Community, Dorazio et al. on multispecies bird communities, PLOS one.

Building statistical models to analyze species distributions, Latimer et al. on hierarchical SDMs, Ecol Appl.

Generalized joint attribute modeling for biodiversity analysis, Clark et al. on combining data types, Ecol Monogr.

11. Space and time

Models for Ecological Data, chapter 9, 10.

Rate of species invasion

Hierarchical Bayesian models for predicting the spread of ecological processes, Wikle on spatial-state-space modeling for invasive species, Ecology.

How does climate change impact phenology?

Divergent responses to spring and winter warming drive community level flowering trends, Cook, Wolkovich, Parmesan find offsetting responses to winter/spring warming, PNAS.

Analyzing First Flowering Event Data using Survival Models with Space and Time-Varying Covariates. Terres et al, as time-to-event, Environmetrics

Tree phenology responses to winter chilling and spring warming, at north and south range limits, Clark et al, as state-space model, Func Ecol.

course information

course information


Nicholas School of the Environment, Department of Statistical Science

office: A201 LSRC

jimclark@duke.edu

TA: Brad Tomasek

needed for class

  • bring a laptop
  • install RStudio

reference

Clark, J.S. 2007. Models for Ecological Data. Princeton University Press. It’s cheap ($32), but you can survive the course without it. You don’t need the lab manual–we’ll use new examples.

objectives

  • recognize:
    • types of samples (observational, experimental)
    • types of observations
    • how data types map to model types
    • limitations of all of the above
  • comprehension:
    • basic distribution theory for connecting data and models
    • foundational differences between traditional/Bayes/machine learning
    • model assumptions
    • diagnostics
    • variable selection
    • analysis to decision
  • implementation:
    • effective use of R
    • acquiring, visualizing, summarizing data sets
    • question to data to model to computation
    • communicating analysis
    • graphical presentation
  • critique model analyses
  • complete a semester project related to your interests

grading

  • participation–show up, contribute in class and to working groups
  • assignments (see below) 3 – great, 2 – ok, 1 – revise
  • final report (see below)
  • final presentation (see below)

structure

  • topical units, each lasting roughly one week
  • readings from the literature
  • reference chapters in Models for Ecological Data
  • a vignette on the course web site corresponds to each unit
    • includes concepts and R code for course
    • includes problem sets: completed in working groups, written up individually
  • class feedback: We are fully accessible for one-on-one interactions; send an email. However, questions about course material should first be posed in class or to your working group. Simple rationale: your questions benefit the entire group.

working groups and assignments

  • groups of 3 to 5 students work on assignments together
  • all members responsible for vignette material
  • designate coordinator for each assignment; rotate
  • agree on what each member should have done before class
  • meetings happen in and outside class
  • review returned assignments; resubmit if necessary
  • peer assessments: confidential feedback
  • late policy: highest possible grade is a 2

semester project

Projects address a problem interest, often focused on graduate research. The final report and presentation informs the class about your project. The report should not exceed 8 pages of text and graphs. Any code goes in an Appendix.

Structure of the report:

  • Introduction
    • the question: what is it, why is it important, potential outcomes (cite literature)
    • the data: type, distribution
    • the model
    • computation
    • status of project
  • Methods
    • plots of the data
    • graphical model
    • data simulation
    • pseudocode
    • diagnostics
  • Interpretation
  • Current Status: what worked, what did not, what did you learn
  • Future directions: do you plan to continue the analysis? Give us an idea of where the project might be headed.
  • Citations
  • Appendix

We will circulate the final reports to reviewers for feedback. Grading will be based on the peer reviews.

gjam

gjam

latest release gjam 2.2.1 on 4-1-18

Generalized Joint Attribute Modeling (GJAM) in R

Ecological attributes include species abundances, traits, and individual condition (e.g., growth or infection status), to name a few. They are multivariate data, but not all of one type.  They can be combinations of presence-absence, ordinal, continuous, discrete, composition, or zero-inflated.   gjam  provides inference on sensitivity to input variables, correlations between responses, model selection, prediction of responses, inverse prediction of predictors, and community classification by response to predictors.

gjam was motivated by species distribution and abundance data, but can provide an attractive alternative to traditional methods wherever observations are multivariate and combine multiple scales and mixtures of continuous and discrete data.

Importantly, analysis is done on the observation scale. That is, coefficients and covariances are interpreted on the same scale as the data. This contrasts with standard generalized linear models (GLMs), where coefficients and covariances are difficult to interpret and cannot be compared across responses that are modeled on different scales and with nonlinear link functions.

 gjam accommodates massive zeros in multivariate data by avoiding the standard mixtures used in zero-inflated GLMs. Instead, gjam relies on censoring.

gjam exploits censoring to combine multiple data types in a single model, including mixtures of continuous and discrete data.  For example, the microbial community (composition data) might be tracked together with host condition (continuous, categorical, binary, ordinal, …).

 

Model:

Clark, J.S., D. Nemergut, B. Seyednasrollah, P. Turner, and S. Zhang. 2017.  Generalized joint attribute modeling for biodiversity analysis: Median-zero, multivariate, multifarious data.   Ecological Monographs, 87, 34–56.  Clark2017EcolMonogrclarksupplement.

Presents the motivation and model; summarizes computation in gjam.  The Supplement file provides additional detail on algorithms.

Dimension reduction: 

Taylor-Rodríguez, D., Kaufeld, K., Schliep, E. M., Clark, J. S., Gelfand, A. E. 2017. Joint species distribution modeling: Dimension reduction using Dirichlet processes. Bayesian Analysis, doi: 10.1214/16-BA1031. http://projecteuclid.org/euclid.ba/1478073617  bayesanaly2016

Many applications require large numbers of response variables.  Microbiome studies bring the additional complication of composition data.  And most observed values can still be zero.  This paper describes the Dirichlet process prior implemented in gjam that finds a low-dimensional representation for the covariance between responses.

Trait analysis:  

Clark, J.S. 2016.  Why species tell us more about traits than traits tell us about species: Predictive models. Ecology,97, 1979–1993, ecology2016ecology2016_AppendixS1

The joint distribution of ecological attributes (‘traits’) can be modeled together with species, separately, or predicted from the joint distribution of species.  This paper describes the model and computation implemented in gjam.

 

Vignette with R code and applications: gjam vignette

Below are cluster plots of the correlation matrix for a presence-absence model (a), continuous abundance model (b), and the response to environmental variables (d).  The cluster analysis in (c) is based on distances in (d).  These plots are obtained by specifying GRIDPLOTS=T in gjamPlot.

fig7a

fig7b

 

 

 

Main contributors:

Jim Clark wrote the GJAM model, the R and C++ code, and the GJAM package.

Alan Gelfand and Daniel Taylor-Rodrigues wrote the Dirichlet process model and algorithms for dimension reduction.

Daniel Taylor-Rodrigues implemented the Dirichlet process in R and C++ in GJAM.

Bene Bachelot, Chase Nuñes, and Brad Tomasek provided extensive testing and feedback through all stages of development.

Many others: Students in the course Bayesian Inference Environm Models (BIO/ENV 665) at Duke University and members of the Multivariate Modeling working group of the SAMSI Ecology program contributed many ideas, recommendations, and feedback.

 

Installation in R or RStudio:

> install.packages('gjam')
> library('gjam')

Documentation:

> help('gjam')
> browseVignettes('gjam')

Publications using GJAM

Bachelot B., Uriarte M., Muscarella R., Forero-Montana J., Thompson J., McGuire K., Zimmerman J.K., Swenson N.G. and J.S. Clark. 2017. Associations among arbuscular mycorrhizal fungi and seedlings are predicted to change with tree successional status. Ecology, in press.

Clark, J.S. 2016.  Why species tell us more about traits than traits tell us about species: Predictive models. Ecology,97, 1979–1993,

Clark, J.S., D. Nemergut, B. Seyednasrollah, P. Turner, and S. Zhang. 2017.  Generalized joint attribute modeling for biodiversity analysis: Median-zero, multivariate, multifarious data.   Ecological Monographs, 87, 34–56.

Taylor-Rodríguez, D., Kaufeld, K., Schliep, E. M., Clark, J. S., Gelfand, A. E. 2017. Joint species distribution modeling: Dimension reduction using Dirichlet processes. Bayesian Analysis, doi: 10.1214/16-BA1031. http://projecteuclid.org/euclid.ba/1478073617  

 

Long-term forest demography

Long-term forest demography

Project PI

  • Jim Clark, Nicholas School of the Environment, Duke University

USFS collaborators

  • Chelcy Ford Miniat
  • Jim Vose

Postdoctoral associates, past and present

  • Sean McMahon, SERC
  • Jessica Metcalf, Princeton University
  • Soledad Benetiz Ponce, Duke University
  • Wei Wu, Assistant Professor, Univ Southern Mississippi

PhD students, past and present

  • Brian Beckage, Univ Vermont
  • Dave Bell,Univ Wyoming
  • Aaron Berdanier, Duke
  • Mike Dietze, Boston University
  •  Michelle Hersh, Eastern Michigan State
  • Janneke HilleRisLambers, Univ Washington
  • Ines Ibanez,Univ Michigan
  • Matthew Kwit, Duke
  • Shannon LaDeau, Cary Inst
  • Jason McLachlan, Notre Dame
  • Jacqueline Mohan,Univ Georgia
  • Emily Moran, UC Merced
  • Brad Tomasek, Duke
  • Mike Wolosin, Pew Center for Climate Change
  • Pete Wyckoff, Univ Minnesota, Morris
  • Kai Zhu, Nicholas School of the Environment, Duke University

Funding: National Science Foundation

Synopsis

The Long-term Forest Demographic (LTFD) Analysis was established to understand how climate and competition interact to control change in eastern forests.  The network, originally established as five stands in the southern Appalachians in 1991, has expanded to include the Piedmont and foothills and now eastern North and Central America.  The project has been supported by a number of NSF grants, including the Coweeta LTER.  It continues with support that includes  NSF’s Macrosystems Biology Program.

From the inception observations on natural variation in space and time were combined with experimental manipulation of the competitive environment.  The data set now extends over 20 yrs from 40,000 trees and > 350,000 tree-years.  We have quantified the interactions between temperature, drought, and competition for light and moisture vary widely between species and size classes.

Early studies quantified basic demographic rates, introducing the emerging hierarchical Bayes paradigm, organized as state-space models.  New modeling innovations have continued to play a large role in this research.

Recent publications

  • Bugalho, M.N., I. Ibánez, and J.S Clark. 2013. The effects of deer herbivory and forest type on tree recruitment vary with plant growth stage. Forest Ecology and Management, in press.
  • Clark, J.S., D. M Bell, M. Kwit, A. Powell, And K. Zhu. 2013. Dynamic inverse prediction and sensitivity analysis with high-dimensional responses: application to climate-change vulnerability of biodiversity.  Journal of Biological, Environmental, and Agricultural Statistics, in press.
  • Ghosh, S., D. M. Bell, J.S. Clark, A.E. Gelfand, and P. Flikkema.  2013. Process modeling for soil moisture using sensor network data. Statistical Science, in press.
  • Wu, W., Clark, J.S., and J. Vose. 2013. Response of hydrology to climate change in the southern Appalachian Mountains using Bayesian inference. Hydrologic Processes, in press.
  • Gelfand, A.E., S. Ghosh and J. S. Clark. 2013. Scaling integral projection models for analyzing size demography. Statistical Science, in press.
  • Ward, E.J., D.M. Bell, J.S. Clark and R. Oren. 2012. Hydraulic time constants for transpiration of loblolly pine at Duke FACE. Tree Physiology, 33, 123-134.
  • Ward, E.J., R. Oren, D.M. Bell, J.S. Clark, H.R. McCarthy, H. Seok-Kim and J.-C. Domec. 2012. The effects of long-term elevated CO2 and nitrogen fertilization on stomatal conductance estimated from scaled sap flux measurements at Duke FACE.  Tree Physiology, 33, 135-151.
  • Rapp, J.M., M. R. Silman, J. S. Clark, C.A. J. Girardin, D. Galiano, and R. Tito. 2012. Intra- and inter-specific tree growth across a long altitudinal gradient in the Peruvian Andes. Ecology, 93:2061-2072.
  • Clark, J.S., B.D. Soltoff, A.S. Powell, and Q.D. Read. 2012. Evidence from individual inference for high-dimensional coexistence: long term experiments on recruitment response. PLoS One, 7 e30050. doi:10.1371/journal.pone.0030050.
  • Moran, E.V. and J.S. Clark. 2012. Causes and consequences of unequal seed production in forest trees: a case study in red oaks. Ecology, 93:1082-1094.
  • Ghosh, S., A.E. Gelfand, K. Zhu, and J.S. Clark. 2012. The k-ZIG: flexible modeling for zero-inflated counts. Biometrics, 68:878-85.
  • Moran, E.V., J. Willis, and J.S. Clark. 2012. Genetic evidence for hybridization in red oaks. American Journal of Botany, 99, 92-100.
  • Clark, J.S., D. M. Bell, M. Kwit, A. Powell, R. Roper, A. Stine, B. Vierra, and K. Zhu. 2012. Individual‐scale inference to anticipate climate‐change vulnerability of biodiversity. Philosophical Transactions of the Royal Society B, 367, 236-246.
  • Uriarte M., J. S. Clark, J. K. Zimmerman, L. S. Comita, J. Forero-Montaña, and J. Thompson. 2012. Multi-dimensional tradeoffs in species responses to disturbance: Implications for diversity in a subtropical forest. Ecology, 93:191–205.
  • Ghosh, S., A. E. Gelfand, and J. S. Clark. 2012.  Inference for size demography from point pattern data using integral projection models. Journal of Agricultural, Biological and Environmental Statistics, 17, 641-677.
  • Zhu, K., C.W. Woodall, and J.S. Clark. 2012. Failure to migrate: lack of tree range expansion in response to climate change. Global Change Biology, DOI: 10.1111/j.1365-2486.2011.02571.
  • Wu, W., Clark, J.S., and J.M. Vose. 2012. Application of a full hierarchical Bayesian model in assessing streamflow response to a climate change scenario at the Coweeta Basin, NC, USA.  Journal of Resources and Ecology, 3, 118-128.
  • PLoS ONE 6(11): e27462. doi:10.1371/journal.pone.0027462
  • Colchero, F. and J.S. Clark. 2011. Bayesian inference on age-specific survival for censored and truncated data. Journal of Animal Ecology 80 DOI: 10.1111/j.1365-2656.2011.01898.x
  • Clark, J.S., D.M. Bell, M.H. Hersh, M. Kwit, E. Moran, C. Salk, A. Stine, D. Valle, and K. Zhu. 2011. Individual-scale variation, species-scale differences: inference needed to understand diversity.  Ecology Letters 14, 1273–1287.
  • Luo,Y. K. Ogle, C. Tucker, S. Fei, C, Gao, S. Ladeau, J. S. Clark, and D. S. Schimel. 2011. Ecological forecasting and data assimilation in a data-rich era. Ecological Applications 21, 1429–1442.
  • Agarwal, P., T. Mohave, H. Yu, and J. S. Clark. 2011. Exploiting temporal coherence in forest dynamics simulation. SCG ’11 Proceedings of the 27th Annual Symposium on Computational Geometry, Paris, France.
  • Clark, J.S., P. Agarwal, D.M. Bell , P. Flikkema , A. E. Gelfand, X. Nguyen , E. Ward, and J. Yang. 2011. Inferential ecosystem models, from network data to prediction. Ecological Applications, 21,1523–1536.
  • Luo Y.Q., J. Melillo, S.L. Niu, C. Beier, J.S. Clark, A.T. Classen, E. Davidson, J.S. Dukes, R.D. Evans, C.B. Field, C.I. Czimczik, M. Keller, B.A. Kimball, L. Kueppers, R.J. Norby, S.L. Pelini, E. Pendall, E. Rastetter, J. Six, M. Smith, M. Tjoelker, M. Torn. 2011. Coordinated approaches to quantify long-term ecosystem dynamics in response to global change. Global Change Biology, 17, 843-854, DOI: 10.1111/j.1365-2486.2010.02265.x.
  • Clark, J.S., D.M. Bell, M.H. Hersh, and L. Nichols. 2011. Climate change vulnerability of forest biodiversity: climate and resource tracking of demographic rates. Global Change Biology, 17, 1834–1849.
  • Wu, W., J.S. Clark, and J. Vose. 2010. Assimilating multi-source uncertainties of a parsimonious conceptual hydrological model using hierarchical Bayesian modeling, Journal of Hydrology, 394, 436-446.
  • Moran, E.V. and J.S. Clark. 2010. Estimating seed and pollen movement in a monoecious plant: a hierarchical Bayesian approach integrating genetic and ecological data. Molecular Ecology, 20, 1248–1262.
  • Clark, J.S., D. Bell, C. Chu, B. Courbaud, M. Dietze, M. Hersh, J. HilleRisLambers, I. Ibanez, S. L. LaDeau, S. M. McMahon, C.J.E. Metcalf, J. Mohan, E. Moran, L. Pangle, S. Pearson, C. Salk, Z. Shen, D. Valle, and P. Wyckoff. 2010. High dimensional coexistence based on individual variation: a synthesis of evidence. Ecological Monographs, 80, 569–608.
  • Clark, J.S. 2010. Individuals and the variation needed for high species diversity. Science 327:1129-1132.
  • Vieilledent, G., B. Courbaud, G. Kunstler, J.-F. Dhote, and J.S. Clark. 2010. Individual variability in tree allometry determines light resource allocation in forest ecosystems: a hierarchical Bayesian approach. Oecolgia, 163: 759-773.
  • Clark, J.S., D. Bell, M. Dietze, M. Hersh, I. Ibanez, S. LaDeau, S. M. McMahon, C.J.E. Metcalf, E. Moran, L. Pangle, and M. Wolosin. 2010. Models for demography of plant populations.  Pages 431 – 481 in T. O’Hagan and M. West (eds) Handbook of Bayesian Analysis, Oxford University Press.
Maps

Maps

Click on any portion of the map to learn more about the work done in our field sites by the Clark Lab.

BW DF NC CW CW MH GSM CL

 

Construction Photos – 2008

Construction Photos – 2008

Harvard Forest Site . . .

. . . Duke Forest Site


Positions

Positions

No Current Openings

 

Wireless Sensor Data

Wireless Sensor Data

Data from the wireless sensor network at Duke Forest, include PAR, air and soil temperature, soil moisture, and weather.  The metadata file includes descriptions of sites and data columns.

EnoEast

EnoEast_wisard_light_2006                                        EnoEast_wisard_light_2007

EnoEast_wisard_light_2008                                        EnoEast_wisard_light_2009

EnoEast_wisard_light_2010                                        EnoEast_wisard_light_2011

EnoEast_wisard_light_2012                                        EnoEast_wizard_light_2013

 

EnoEast_wisard_temp_2006                                       EnoEast_wisard_temp_2007

EnoEast_wisard_temp_2008                                       EnoEast_wisard_temp_2009

EnoEast_wisard_temp_2010                                       EnoEast_wisard_temp_2011

EnoEast_wisard_temp_2012

EnoEast_wizard_soilTemp_2013

EnoEast_wizard_airTemp_2013

 

EnoEast_wisard_soilmoisture_2006                           EnoEast_wisard_soilmoisture_2007

EnoEast_wisard_soilmoisture_2008                           EnoEast_wisard_soilmoisture_2009

EnoEast_wisard_soilmoisture_2010                           EnoEast_wisard_soilmoisture_2011

EnoEast_wisard_soilmoisture_2012                          EnoEast_wizard_soilmoisture_2013

 

EnoWest

EnoWest_wisard_light_2006                                        EnoWest_wisard_light_2007

EnoWest_wisard_light_2008                                        EnoWest_wisard_light_2009

EnoWest_wisard_light_2010                                        EnoWest_wisard_light_2011

EnoWest_wisard_light_2012                                        EnoWest_wizard_light_2013

 

EnoWest_wisard_temp_2006                                       EnoWest_wisard_temp_2007

EnoWest_wisard_temp_2008                                       EnoWest_wisard_temp_2009

EnoWest_wisard_temp_2010                                       EnoWest_wisard_temp_2011

EnoWest_wisard_temp_2012

EnoWest_wizard_soilTemp_2013

EnoWest_wizard_airTemp_2013

 

EnoWest_wisard_soilmoisture_2006                            EnoWest_wisard_soilmoisture_2007

EnoWest_wisard_soilmoisture_2008                            EnoWest_wisard_soilmoisture_2009

EnoWest_wisard_soilmoisture_2010                            EnoWest_wisard_soilmoisture_2011

EnoWest_wisard_soilmoisture_2012                            EnoWest_wizard_soilmoisture_2013

 

EnoWest_wisard_airpres

EnoWest_wisard_rainaccu

EnoWest_wisard_relhum

EnoWest_wisard_winddir

EnoWest_wisard_windspeed

 

Hardwood

Hardwood_wisard_light_2006                                      Hardwood_wisard_light_2007

Hardwood_wisard_light_2008                                      Hardwood_wisard_light_2009

Hardwood_wisard_light_2010                                      Hardwood_wisard_light_2011

Hardwood_wisard_light_2012                                      Hardwood_wizard_light_2013

 

Hardwood_wisard_soilmoisture_2006                         Hardwood_wisard_soilmoisture_2007

Hardwood_wisard_soilmoisture_2008                         Hardwood_wisard_soilmoisture_2009

Hardwood_wisard_soilmoisture_2010                         Hardwood_wisard_soilmoisture_2011

Hardwood_wisard_soilmoisture_2012                         Hardwood_wizard_soilmoisture_2013

 

Hardwood_wisard_temp_2006                                     Hardwood_wisard_temp_2007

Hardwood_wisard_temp_2008                                     Hardwood_wisard_temp_2009

Hardwood_wisard_temp_2010                                     Hardwood_wisard_temp_2011

Hardwood_wisard_temp_2012

Hardwood_wizard_soilTemp_2013

Hardwood_wizards_airTemp_2013

 

META file

META_WiSARD_Definitions

(more…)

Ecosystem sensing

Ecosystem sensing

Large-Scale Wireless Sensor Networks for Observation of Ecosystem Processes

Funding: NSF IDEA-0308498

Progress in an array of technologies, including microelectronic sensing and computation, wireless communication, and the self-assembly of autonomous devices into cooperative networks has inspired the vision of wireless sensor networks. While networks of intelligent agents transparently embedded into our physical environment could advance human welfare in a number of domains, research indicates that any successful wireless sensor network must be carefully optimized for its application. One of the most compelling of these applications is dense spatio-temporal sensing of environments to enable better understanding of environmental and ecosystem processes across multiple scales. Our goals are to (i) test this hypothesis in three rigorous field studies, (ii) bootstrap the application of wireless sensor network technology in many applications, and (iii) build awareness of the benefits of the technology to society, and improve collaboration between engineering and the sciences.

The instrumentation development component of this project builds on a successful seed effort in which we have constructed a small proof-of-concept wireless environmental sensing network. We will build a “distributed instrument”—a prototype network comprising hundreds of palm-sized wireless sensors. In anticipation of this prototype development, we have paid careful attention to ensuring that our networking technology would successfully scale up to hundreds and thousands of sensors at up to landscape geographic scales.

The prototype network technology will be deployed to enable a new degree of data quality in three diverse field studies. First, will probe the role of the role of fine-scale environmental phenomena in the maintenance of ecosystem diversity in two Eastern US forests. The second experiment maps the complexity of microclimates in the crowns of the coastal redwoods of California. And in the third field study, we will determine the effects of scale on eddy covariance measurements of ecosystem energy balance in Northern Arizona.

Collaborators: Pankaj Agarwal, Carla Ellis, Paul Flikkema, Alan Gelfand, Kamesh Munagala, Jun Yang

Recent publications:

  • Ghosh, S., D. M. Bell, J.S. Clark, A.E. Gelfand, and P. Flikkema.  2013. Process modeling for soil moisture using sensor network data. Statistical Science, in press.
  • Wu, W., Clark, J.S., and J. Vose. 2013. Response of hydrology to climate change in the southern Appalachian Mountains using Bayesian inference. Hydrologic Processes, in press.
  • Ward, E.J., D.M. Bell, J.S. Clark and R. Oren. 2012. Hydraulic time constants for transpiration of loblolly pine at Duke FACE. Tree Physiology, 33, 123-134.Ward, E.J., R. Oren, D.M. Bell, J.S. Clark, H.R. McCarthy, H. Seok-Kim and J.-C. Domec. 2012. The effects of long-term elevated CO2 and nitrogen fertilization on stomatal conductance estimated from scaled sap flux measurements at Duke FACE.  Tree Physiology, 33, 135-151.
  • Hersh, M.H., J.S. Clark, and R. Vilgalys. 2012. Evaluating the impacts of fungal seedling pathogens on temperate forest seedling survival. Ecology, 93: 511-520.
  • Wu, W., Clark, J.S., and J.M. Vose. 2012. Application of a full hierarchical Bayesian model in assessing streamflow response to a climate change scenario at the Coweeta Basin, NC, USA.  Journal of Resources and Ecology, 3, 118-128.
  • Agarwal, P., T. Mohave, H. Yu, and J. S. Clark. 2011. Exploiting temporal coherence in forest dynamics simulation. SCG ’11 Proceedings of the 27th Annual Symposium on Computational Geometry, Paris, France.
  • Clark, J.S., P. Agarwal, D.M. Bell , P. Flikkema , A. E. Gelfand, X. Nguyen , E. Ward, and J. Yang. 2011. Inferential ecosystem models, from network data to prediction. Ecological Applications, 21,1523–1536.
  • Wu, W., J.S. Clark, and J. Vose. 2010. Assimilating multi-source uncertainties of a parsimonious conceptual hydrological model using hierarchical Bayesian modeling, Journal of Hydrology, 394, 436-446.
  • Flikkema, P.G., P.J. K. Agarwal, J. S. Clark, C. Ellis, A. Gelfand, K. Munagala, and J. Yang. 2007. From data reverence to data relevance:  Model-mediated wireless sensing of the physical environment. Pages 988–994 in Y. Shi et al. (Eds.): ICCS 2007, Part I, LNCS 4487.
  • Flikkema, P.G., P.K. Agarwal, J.S. Clark, C. Ellis, A. Gelfand, K. Munagala, and J. Yang.  2006. Model-driven dynamic control of embedded wireless sensor networks.  Proc. 6th International Conference on Computational Science, Workshop on Dynamic Data Driven Application Systems, Reading, UK
Data Sets

Data Sets

The NEON data

NEON data are available from the NEON data portal by sample date and subplot.  In cases where data on plants, ground beetles, and small mammals are desired in more aggregate form it can be expedient to use the NEON data aggregator.

 

The Warming Experiment

The Harvard Forest (near Petersham, Worcester County, Massachusetts) and Duke Forest (near Hillsborough, Orange County, North Carolina) warming experiment sites share the same experiment design and instrumentation.  At each site, soil and air heating is carried out in eighteen open-topped chambers and 6 non-chamber controls.  Experimental treatments include three levels of heating (ambient/control, +3° C, and +5° C) and low-light (understory) versus high-light (artificially created gap) conditions.  Each of these chambers contains sensors to measure air temperature, soil temperature, relative humidity, soil moisture, and photosynthetically active radiation.  These environmental data, recorded on an hourly basis by Campbell Scientific CR1000 data loggers, are available below.


Duke Forest Environmental Data

Duke Forest Metadata including experiment treatments, column definitions, and important dates.

Duke Forest Air Temperature

Duke Forest Soil Temperature

Duke Forest Soil Moisture

Duke Forest Photosynthetically Active Radiation

Relative humidity not currently available.

 

Harvard Forest Environmental Data

Harvard Forest Metadata including experiment treatments, column definitions, and important dates.

Harvard Forest Air Temperature

Harvard Forest Soil Temperature

Harvard Forest Soil Moisture

Harvard Forest Photosynthetically Active Radiation

Relative humidity not currently available.

 

Duke Forest Spring Phenology Data

Gap Chambers

G01_A

G02_5

G03_3

G04_A

G05_3

G06_5

G07_A

G08_5

G09_3

G10_C

G11_C

G12_C

Shade Chambers

S01_5

S02_3

S03_A

S04_A

S05_3

S06_5

S07_5

S08_A

S09_3

S10_C

S11_C

S12_C

Spring Phenology Table

 

Harvard Forest Spring Phenology Data

Gap Chambers

G01_3

G02_A

G03_5

G04_A

G05_5

G06_3

G07_A

G08_3

G09_5

G10_C

G11_C

G12_C

Shade Chambers

S01_5

S02_A

S03_3

S04_5

S05_A

S06_3

S07_A

S08_3

S09_5

S10_C

S11_C

S12_C

Contact: Jordan Siminitz (jordan.siminitz@duke.edu)

Last updated: June 3rd, 2014

Forest biodiversity and climate at the macro scale

Forest biodiversity and climate at the macro scale

 

Link to project web site

 

Over two decades of experimental research and modeling now shifts to eastern North America.  The collaboration is funded by the Macrosystems program of the National Science Foundation and includes:

Jim Clark – Duke University

Mike Dietze – University of Illinois

Andrew Finley – Michigan State University

Alan Gelfand – Duke University

Sean McMahon – Smithsonian

Jackie Mohan – University of Georgia

Maria Uriarte – Columbia University

Climate change is rapidly transforming forests over much of the globe in ways that are not anticipated by current science.  Large-scale forest diebacks, apparently linked to interactions involving drought, warm winters, and other species, are becoming alarmingly frequent.  Models of biodiversity and climate have not provided guidance on if/where/when such responses will occur.  Instead models often predict potential numbers of extinctions, but these forecasts not are linked in any mechanistic way to the processes that could cause them.  Both modeling and field studies rely on aggregate metrics of species presence/absence or relative abundance at regional scales, but climate affects individuals.  Aggregation of individual data to the species level, hides or even qualitatively changes climate effects.  By sampling and analysis at the individual scale across continental variation in climate, this study can link the individual scale processes to regional responses.  This study will exploit existing research sites and the new NEON platform of sites for synthesis of models and data to determine when and where predicting climate impacts on biodiversity is a plausible goal, understand where surprises are likely to occur, and attribute those predictions back to individual tree health and vulnerability to climate risk factors.

The study will provide climate vulnerability forecasts for forest biodiversity that are directly linked to the process scale.  Our goal is provide probabilistic forecasts for the joint distribution of forest responses to climate change, including growth, reproduction, and mortality risk.  For scientists, US Forest Service researchers, and policy makers predictions will anticipate combined risks of increasing drought and longer growing seasons.  Methods developed under this project will be disseminated through training workshops for postdoctoral associates at other universities and resource managers.

All sites have been implemented as of 7/2013

 

Recent Publications

  • Clark, J.S., A.E. Gelfand, C.W. Woodall, and K. Zhu. 2013. More than the sum of the parts: forest climate response from joint species distribution models, Ecological Applications, in press.
  • Clark, J.S., D.M. Bell, M.C. Kwit, and K. Zhu. 2013. Competition-interaction landscapes for the joint response of forests to climate change. Global Change Biology, in press.
  •  Zhu, K, C. W. Woodall, S. Ghosh, A. E. Gelfand, and J. S. Clark. 2013. Dual impacts of climate change: forest migration and turnover through life history. Global Change Biology, in press.
  • Clark, J.S., D. M Bell, M. Kwit, A. Powell, And K. Zhu. 2013. Dynamic inverse prediction and sensitivity analysis with high-dimensional responses: application to climate-change vulnerability of biodiversity.  Journal of Biological, Environmental, and Agricultural Statistics, 18:376-404.
Courses

Courses

Bayesian Inference for Environmental Models (BIO/ENV 665, Fall 2018)

Jim Clark

Meetings: MW – 01:25PM to 02:40PM, LSRC A155

prerequisite: one semester each of stats, calculus

 previous years:

data files from lab manual

Environmental Change in the Big Data Era (ENV 89S, Fall 2018)

Jim Clark

Meetings: TTh – 01:25PM to 02:40PM, ENV1101

Generalized Joint Attribute Modeling (GJAM) at NEON

University of Tennessee, 25 Jan 2018

Uncertainty quantification for NEON and other biodiversity network data with hierarchical Bayes

text and code: gjamNeon

 

Ecological diversity and climate change (ENV 623L, Fall 2017)

Jim Clark

Meetings: MW – 11:45AM to 01:00PM in Environment Hall 2102

Lab: W – 01:25PM to 02:40PM in Environment Hall 1104

Site under construction, info from 2016:

BiodiverClimClass

Ecological perspectives: individuals to communities (UPE 701, Fall 2017)

with Susan Alberts

Meetings: TTh 10:05-11:20

previous years:

data files from lab manual

 

Data synthesis workshop: ILTER 1st Open Sci Meeting, Uncertainty Quantification for NEON,

Kruger National Park, 11 Oct 2016

Uncertainty quantification for NEON and other biodiversity network data with hierarchical Bayes

text and code: ilter2016

 

 

Computational advances for forests and climate change

Computational advances for forests and climate change

CDI-Type II: Integrating Algorithmic and Stochastic Modeling Techniques for Environmental Prediction

Project PIs

  • Jim Clark, Nicholas School of the Environment, Duke University
  • Alan Gelfand, Department of Statistical Science, Duke University
  • Pankaj Agarwal, Department of Computer Science, Duke University

USFS collaborator

  • Chris Woodall

Postdoctoral associates

  • Souparno Ghosh, Department of Statistical Science, Duke University
  • Thomas Mølhave, Department of Computer Science, Duke University
  • Joao Montiero, Department of Statistical Science, Duke University

PhD students

  • Maria Terres, Department of Statistical Science, Duke University
  • Kai Zhu, Nicholas School of the Environment, Duke University

Funding: National Science Foundation

Synopsis

Predicting biodiversity, i.e., abundance of species, in response to climate change is a goal of environmental change research. Despite recent valuable advances in understanding biodiversity and climate, the current grasp is limited. There are two widely recognized obstacles: first, because of the complexity of the underlying processes, the existing models intended for understanding and prediction are not (computationally) scalable. Second, the coarse-scale environment models fail to capture interactions among species, which control biodiversity, and the models based on fine-scale, short-term observations are unable to make long-term predictions. This research develops a prediction framework that coherently combines broad-scale pattern data with fine-scale data on species interactions and that is computationally scalable. It focuses on prediction at the geographic scale and in using geographic-scale data to better understanding at the scales where species interactions occur.

This research develops a multiscale modeling framework and design algorithms that make environmental models computationally scalable. The approach hinges upon strong interplay of algorithmic and statistical techniques. Statistical inference brings stochastic modeling sophistication in space and time, yielding improved characterization of the process and the possibility of full inference. Sophisticated algorithms makes models and processes scalable and provide trade-offs between accuracy and efficiency. The project team is composed of researchers from computer science, statistics, and environmental science, who have made significant contributions to these and related problems and who have collaborated extensively in the past.

Recent publications

  • Clark, J.S., A.E. Gelfand, C.W. Woodall, and K. Zhu. 2013. More than the sum of the parts: forest climate response from joint species distribution models, Ecological Applications, in press.
  • Clark, J.S., D.M. Bell, M.C. Kwit, and K. Zhu. 2013. Competition-interaction landscapes for the joint response of forests to climate change. Global Change Biology, in press.
  • Zhu, K, C. W. Woodall, S. Ghosh, A. E. Gelfand, and J. S. Clark. 2013. Dual impacts of climate change: forest migration and turnover through life history. Global Change Biology, in press.
  • Clark, J.S., D. M Bell, M. Kwit, A. Powell, And K. Zhu. 2013. Dynamic inverse prediction and sensitivity analysis with high-dimensional responses: application to climate-change vulnerability of biodiversity.  Journal of Biological, Environmental, and Agricultural Statistics, 18:376-404.
  • Gelfand, A.E., S. Ghosh and J. S. Clark. 2013. Scaling integral projection models for analyzing size demography. Statistical Science, in press.
  • Zhu, K., C.W. Woodall, and J.S. Clark. 2012. Failure to migrate: lack of tree range expansion in response to climate change. Global Change Biology, 18, 1042-1052.
  • Ghosh, S., A.E. Gelfand, K. Zhu, and J.S. Clark. 2011. The k-ZIG: flexible modeling for zero-inflated counts. Biometrics, 68, 878-885.
  • Ghosh, S., A. E. Gelfand, and J. S. Clark. 2011.  Inference for size demography from point pattern data using integral projection models. Journal of Agricultural, Biological and Environmental Statistics, in press.
  • Agarwal, P., T. Mohave, H. Yu, and J. S. Clark. 2011. Exploiting temporal coherence in forest dynamics simulation. SCG’11 Symposium on Computational Geometry, June 13–15, 2011, Paris, France.
  • Clark, J.S., P. Agarwal, D.M. Bell , P. Flikkema , A. E. Gelfand, X. Nguyen , E. Ward, and J. Yang. 2011. Inferential ecosystem models, from network data to prediction. Ecological Applications, 21,1523–1536.
News

News

Most downloaded paper in Global Change Biology
From GCB: Congratulations on your most most downloaded article in Global Change Biology (GCB), “The impacts of increasing drought on forest dynamics, structure, and biodiversity.”  Your paper is one of 2016’s 15 most-downloaded, according to Web of Science®.

Although a relatively young journal, our Impact Factor has been growing steadily and in 2015 the Impact Factor was 8.444 (ISI Journal Citation Reports).  GCB’s 2015 ISI ranking is 1st of 49 in biodiversity conservation, 4th of 225 in Environmental Sciences, 6th of 150 in Ecology Sciences.  These rankings are all the more impressive given that unlike our competitors, we carry very few review articles- which tend to skew Impact Factors upwards.

 

National Drought Assessment honored with USFS Chief’s Award

On Thursday, December 8, Jim Vose, project leader of the U.S. Forest Service Integrated Forest Science accepted the Chief’s Award – one of the highest honors in the Forest Service — in the category of “Sustaining Forests and Grasslands.” Vose accepted as leader of a team that the award honored for “understanding the impacts of drought on the nation’s forests and grasslands: providing a scientific foundation for effective management responses.” read more…

The National Assessment was led by Jim Vose, Jim Clark, Charlie Luce, and Toral Patel-Weynand:  Effects of Drought on Forests and Rangelands in the United States: A Comprehensive Science Synthesis,

 

Int Soc Bayesian Analysis best poster award

The EnviBayes section of ISBA this year has granted two best posters awards at the ISBA World Conference in Forte Village (June 13th – 17th, Cagliari, Italy).  Joint Species distribution modeling: dimension reduction using Dirichlet processes by Daniel Taylor-Rodriguez, postdoc at the Department of Statistical Science Duke University, with Kimberly Kaufield from North Carolina State University, Erin Schliep of University of Missouri, James Clark and Alan Gelfand of Duke University.

The paper is accepted at Bayesian Analysisbayesanaly2016.

Drought synthesis in Global Change Biology posted 22 Feb 2016

Forests nationwide are feeling the heat from increasing drought and climate change, according to a new study by scientists from 14 research institutions.

“Over the last two decades, warming temperatures and variable precipitation have increased the severity of forest droughts across much of the continental United States,” said James Clark, lead author of the study and an environmental scientist at Duke University…

from the National Science Foundation

read the paper: gcb2016

 

The National Drought Assessment released 2 Feb 2016

This assessment provides input to the reauthorized National Integrated Drought Information System (NIDIS) and the National Climate Assessment (NCA); it also establishes the scientific foundation needed to manage for drought resilience and adaptation. The NIDIS Act1 was signed into law in 2006 and reauthorized by Congress in 2014.2 NIDIS will be implemented through a network of agencies
and partners to integrate drought monitoring and forecasting systems
at multiple levels (Federal, State, and local). It will support research
that focuses on drought risk assessment, forecasting, and monitoring. Produced every 4 years, the NCA evaluates the effects of global climate change on forests, agriculture, rangelands, land and water resources, human health and welfare, and biological diversity, and it projects major trends. The NCA is based on technical information produced by public agencies and nongovernmental organizations.

Zhu et al. in Global Change Biology

Many climate studies have predicted that tree species will respond to global warming by migrating via seed dispersal to cooler climates. But a new study of 65 different species in 31 eastern states finds evidence of a different, unexpected response…[more]