B43B-1159
Predictability and detectability of biogeographic changes in plant distributions
Rapid climatic changes are envisioned as atmospheric composition is changed by human activities. These changes have long been predicted to drive large-scale changes in the distribution of plants and of all their associated biota. The direct effects of increasing atmospheric CO2 on photosynthesis, transpiration, and nutrient dynamics have also been predicted to alter the abundance and density of whole functional groups of plants, particularly those differing in photosynthetic pathways (increases in C3 plants at the expense of C4s, as one considerable simplification). In recent work, I have pointed out major physiological diversity among individual plant species in their direct responses to elevated CO2. The consequences include considerable fragmentation in migration patterns of plant species over decades to centuries. Refining the predictions is a daunting task largely in the areas of physiology, ecology, and evolution. Detecting the changes for validation of predictions and for management/ response strategies is similarly a major challenge. Many changes in plant performance and distribution driven directly by climate and CO2 are modest to date, given the modest scale of changes in these two drivers over decadal time scales amenable to both field studies and remote sensing. Large-scale changes, such as in growing season, have occurred but species details have not been resolved in observations with global, repeated coverage. Additional large-scale studies, to merge with small-scale studies, are needed. I review briefly the feasibility of remote-sensing studies for such purposes. For limited campaigns, there is the potential for resolving species by spectral signature, using advance hyperspectral sensing coupled with biophysical models. More general remote-sensing technologies allow detecting shifts between functional types (e.g., grass/woodland) at ecotones, and shifts at all locations in physiological stresses - particularly water stress in its temporal and spatial spectra - directly or via changes in gross primary productivity.
B43B-1160
Remote sensing of threshold conditions in an arid ecosystem
Land management in the arid southwestern USA increasingly addresses thresholds in response to recent concepts adopted by private and public lands agencies and conservation organizations. Vegetation in arid rangelands typically presents as distinctive mosaics of vegetation patches, which persist in dynamic equilibrium with the abiotic environment and facilitative-competitive interactions between organisms. Theory and observation suggest that as an area approaches a threshold in response to disturbance, there is a concomitant change in the spatial arrangement of vegetation patches. This change is readily identifiable on fine spatial resolution aerial photography or satellite sensor imagery. We propose a classification method for identifying threshold-inducing change in vegetation pattern. To illustrate this method, we have applied an object-oriented, supervised classification to subsets of Quickbird imagery (70 cm ground resolution) over the Jornada basin in southern New Mexico. The imagery covers several land management regimes (private, public, federal) and provides spatial variation in ecosystem conditions. Imagery was first segmented to create fine and coarse resolution image objects. Fine resolution image objects are defined as having within-object spectral homogeneity at the scale of the shrub or single patch of grass or soil. Coarse resolution image objects are defined as containing spectral homogeneity at the scale of the vegetation stand. A classification tree was used to classify coarse resolution image objects to high risk of a threshold, low risk of a threshold, or post-threshold according to the content and spatial arrangement of shrub, grass and soil patches within them. Ground-based monitoring to detect localized threshold conditions across broad management areas is intractable so the use of remote sensing is essential to successful prevention of threshold development.
B43B-1161
Ecosystem Performance Anomalies in the Bonanza Creek Area, Alaska
Central Alaska is ecologically sensitive and experiencing stress in response to marked regional warming. We need a better ability to monitor ecosystem processes that are responding to climate change, fire, and insect damage, and to predict responses to future climate and environmental conditions. We have developed a method for analyzing ecosystem performance that illustrates the status and trends of ecosystem changes and that separates the influences of climate and local site conditions from the influences of disturbances and land management practices. The poster shows results of the method via a time series graph of ecosystem performance anomalies for each remotely sensed pixel of a boreal forest area that includes the Bonanza Creek Long Term Ecological Research (LTER) site near Fairbanks, Alaska. Measures of "ecosystem performance" are based on a seasonally integrated normalized difference vegetation index using composited data acquired by NOAA's Advanced Very High Resolution Radiometer (AVHRR). We define an "expected ecosystem performance" to represent the greenness response of vegetation that is expected in a particular year given the climate of that year, and we distinguish "performance anomalies" as cases where the ecosystem response is significantly different than the expected ecosystem performance. This poster illustrates Ecosystem Performance Anomaly Trends (EPAT). The magnitude of the ecosystem performance anomaly is separated into three categories: 1) performing better than expected, 2) performing within the expected range, or 3) performing more poorly than expected. A pixel is classed as anomalously overperforming (or underperforming) if it is above (or below) the 90-percent significance line in 6 of the 8 years modeled. Within each category, we also show if the trend is 1) decreasing, 2) nearly level, or 3) increasing. Combining these dimensions gives nine categories for the map. Recent fires are clearly detected by the method, but other areas of ecosystem stress are also identified.
B43B-1162
Tree- Rings Link Climate and Carbon Storage in a Northern Mixed Hardwood Forest
The terrestrial biosphere is a variable sink for atmospheric carbon dioxide. It is important to understand how carbon storage in trees is affected by natural climate variability to better characterize the sink. Quantifying the sensitivity of forest carbon storage to climate will improve carbon budgets and have implications for forest management practices. Here we explore how climate variability affects the ability of a northern mixed hardwood forest in Michigan to sequester atmospheric carbon dioxide in woody tissues. This site is ideal for studies of carbon sequestration; The University of Michigan Biological Station is an Ameriflux site, and has detailed meteorological and biometric records, as well as CO2 flux data. We have produced an 82- year aspen (Populus grandidentata) tree-ring chronology for this site, and measured ring widths at several heights up the bole. These measurements were used to estimate annual wood volume, which represents carbon allocated to aboveground carbon stores. Standard dendroclimatological techniques are used to identify environmental factors (e.g. temperature or precipitation) that drive tree-ring increment variability in the past century, and therefore annual carbon storage in this forest. Preliminary results show that marker years within the tree- ring chronology correspond with years that have cold spring temperatures. This suggests that trees at this site are temperature sensitive.
B43B-1163 [WITHDRAWN]
Potential mechanisms for fine-scale variation in desertification thresholds
The process of desertification includes many interactions and feedbacks occurring at multiple spatial scales, often resulting in nonlinear dynamics and threshold behavior. These interactions and feedbacks have been observed in the broad-scale transition of former black grama grasslands to mesquite dominated shrublands (desertification) in the Chihuahuan Desert of southern New Mexico, USA. Heterogeneity in soil properties at the landscape scale can increase resistance to change, thereby reducing the probability of crossing a threshold at some sites due to small scale, within patch processes. Observational studies indicate that grass communities occurring in sandy soils shallow to well-developed petrocalcic horizons are more resistant to drought than those growing on similar soils without a this horizon. To assess the mechanism for this observed resistance, water availability dynamics were monitored in a multiyear, landscape scale study across soils with differing degrees of petrocalcic development. Results show that petrocalcic horizons absorbed and retained much greater amounts of available soil water for several months following an extremely wet winter and summer than similar depths in a deep sandy soil. Wetting and drying dynamics indicate the mechanism for the observed resistance of grasses to drought on these soils is the slow release of petrocalcic water into the grass rooting zone. The petrocalcic soil- water dynamics observed during these extreme events indicate that the high water holding capacity of these soils can potentially buffer effects of increased variability in precipitation expected to occur as a result of climate change. In water limited systems, water holding capacity of the entire profile, including rock-like material such as calcium carbonate, should be considered when evaluating the potential resistance of sites to crossing a threshold, particularly in a changing climate. Areas that are inherently more resistant to patch-scale shrub invasion processes could present a management opportunity for restoration or conservation--if they can be protected from broad-scale feedbacks and drivers pushing the system towards a desertification threshold.
B43B-1164
Fine Root and Aboveground Development and Turnover in a Seasonally Dry Forest
Terrestrial carbon dynamics are dependent in large part on factors which affect fine root dynamics, as fine roots constitute a significant part of belowground carbon allocation. Differences in factors affecting fine root dynamics under different climatic conditions are not well known. We conducted a three year study of above- and belowground dynamics in a Ponderosa pine plantation in the central Sierra Nevada, examining the overall patterns of fine root production and survival in a seasonally arid montane forest in a Mediterranean climate, as well as the correlations between above- and belowground growth dynamics. Data collected during the 2003- 2005 seasons suggest that fine root development of ponderosa pine-dominated ecosystems of the Sierra Nevada are highly dependent on prevailing climatic conditions, and that belowground phenology is more sensitive to variations in available moisture than is aboveground development. To the best of our knowledge, these findings represent one of the first comprehensive examinations of above- and below-ground growth dynamics in a forest system growing in a Mediterranean climatic zone, and offer a few clues for evaluating the potential difficulties in projecting climate change influences on ecosystem dynamics. We monitored stem diameter growth, shoot and needle elongation, fine root development using minirhizotrons, as well as soil moisture, and soil and air temperature throughout the season. The onset of belowground phenological development in this ecosystem appears to be controlled by the increase of the daily minimum temperatures above 5 ° C, and the end of the active fine root elongation season appears to be controlled by the decrease of soil water potential below -0.3MPa. Root development appears to follow the overall patterns of warming and drying of the soil profile, with surface root growth initiation and mortality occurring earlier than that of subsurface roots. In contrast, aboveground development does not appear to be strongly related to changes in climatic conditions.
B43B-1165
Predicting Coral Bleaching: Remote Sensing vs. In Situ Measurements of Sea Surface Temperatures
Here we present a case study where we compare the skill of predicting coral bleaching using two sources of Sea Surface Temperature (SST) data. Remotely sensed microwave (TMI) data and in situ observations of SST are used to calculate Degree Heating Weeks, a NOAA technique to predict thermal stress on coral reefs. The in situ observations have been made over a period of 1993-2007 at the Le Port, Reunion. We calculate DHW and compare it to a database of reef observations at Reunion, enabling us to assess skill for the predictive method using both sources of SST data. Quantitative accuracy assessment, using well established methods and skill scores often used in meteorology and medical sciences (Peirce Skill Score, Hit Rate and False Alarm Rate), enable objective optimization of the predictive technique for each SST source. The differences in skill expressed in Hit Rate and False Alarm Rate have implications for decision making concerning reef management and conservation techniques.
B43B-1166
A method to objectively optimize coral bleaching prediction techniques
Thermally induced coral bleaching is a global threat to coral reef health. Methodologies, e.g. the Degree Heating Week technique, have been developed to predict bleaching induced by thermal stress by utilizing remotely sensed sea surface temperature (SST) observations. These techniques can be used as a management tool for Marine Protected Areas (MPA). Predictions are valuable to decision makers and stakeholders on weekly to monthly time scales and can be employed to build public awareness and support for mitigation. The bleaching problem is only expected to worsen because global warming poses a major threat to coral reef health. Indeed, predictive bleaching methods combined with climate model output have been used to forecast the global demise of coral reef ecosystems within coming decades due to climate change. Accuracy of these predictive techniques has not been quantitatively characterized despite the critical role they play. Assessments have typically been limited, qualitative or anecdotal, or more frequently they are simply unpublished. Quantitative accuracy assessment, using well established methods and skill scores often used in meteorology and medical sciences, will enable objective optimization of existing predictive techniques. To accomplish this, we will use existing remotely sensed data sets of sea surface temperature (AVHRR and TMI), and predictive values from techniques such as the Degree Heating Week method. We will compare these predictive values with observations of coral reef health and calculate applicable skill scores (Peirce Skill Score, Hit Rate and False Alarm Rate). We will (a) quantitatively evaluate the accuracy of existing coral reef bleaching predictive methods against state-of- the-art reef health databases, and (b) present a technique that will objectively optimize the predictive method for any given location. We will illustrate this optimization technique for reefs located in Puerto Rico and the US Virgin Islands.
B43B-1167
Thresholds of Carbon Assimilation and Respiration in a California's Oak/Grass Savanna
Multi-year ecosystem CO2 fluxes measurement provides us opportunities to understand dynamics and thresholds of carbon assimilation and respiration. Based on eddy-covariance-measured ecosystem carbon assimilation and respiration in a California's oak/grass savanna from 2001 to 2006, we examined the thresholds of these two major processes that control ecosystem carbon uptake. Our preliminary results indicated that year- to-year variations in water supplies (i.e., precipitation) significantly controlled interannual variability in carbon uptake in the savanna ecosystem. Seasonally, grass photosynthesis decreased with decreases in soil moisture, but our results suggested multiple degrading peaks, indicating coupling effects of temperature. On the contrary, tree photosynthesis responded to decreases soil moisture with one significant peak. The thresholds of soil moisture at maximum tree photosynthesis were around 0.1 cm cm-3. The lowest soil moisture threshold occurred in the dry year, 2004. Annual grass grew during the period when soil temperature was lower than 25oC, while oak tree was active during the period with soil temperature between 10 to 35oC. Grass reached maximum photosynthesis at 13.1, 13.7, 16.1, 15.4, and 16.9oC in the spring of 2002, 2003, 2004, 2005, and 2006, respectively. Under current climate variation, difference of temperature thresholds of grass photosynthesis spanned up to 3.8oC, and difference of soil moisture thresholds of tree photosynthesis spanned up to 0.11 cm cm-3. On the other hand, thresholds of ecosystem respiration were coupled with those of photosynthesis responding to interannual variations of soil temperature and moisture. Implications of ecosystem thresholds to global change are also to be discussed.
B43B-1168
Climate and soil-age constraints on nutrient uplift by plants.
We analyzed changes in nutrient availability and elemental losses from the entire weathering zone at 28 sites arrayed across climatic and soil-age gradients on the island of Hawai'i. The sites are located on three basaltic lava flows (10, 170, and 350 ky) each of which crosses a precipitation gradient from <500 to 2,500 mm yr- 1. The results identify a sweet spot of plant nutrient uplift where nutrient cations and phosphorus are retained in upper horizons as a result of plant activity. The gradients also elucidate several abiotic constraints on plant- driven retention of nutrients. At the dry sites (<750 mm yr-1on all three flows, plant slow the loss of nutrient (e.g. potassium) vs. non-nutrient (e.g. sodium) cations, but the effect is small because of low plant cover and productivity. At intermediate rainfall (750 - 1300 mm yr-1) plants substantially enrich both nutrient cations and P in the upper soils, an effect that increases with flow age. In contrast, at high rainfall (>1500 mm yr-1), the effect of plants on nutrient distributions diminishes with soil age and is largely absent after 350 ky of soil development. Unlike the major plant macronutrients, the distribution of the transition metals iron (Fe) and aluminum (Al) is driven more by chemical reactions than by plant uptake. Dry sites exhibit very little movement of either element, even after 350 ky of soil development. However at high rainfall the older flows show substantial Al and Fe translocations, and wet sites on all three flows have increased Al on soil exchange sites. These transition metals are key constituents of the secondary minerals that strongly influence the availability of cations and P to plants. The loss of Fe and Al is highly correlated with the loss of P in the older and wetter sites, and increased Al on exchange sites limits the availability of nutrient cations to plants. Thus redox driven redistribution of Fe and acid solublization of Al place a further abiotic constraint on nutrient retention by plants.
B43B-1170
Rainfall, nitrogen deposition and fire disturbance impacts in a California coastal grassland
In semi-arid ecosystems, shifts in soil moisture availability may mediate the response of individual species, communities and ecosystems to disturbance or changes in nutrient availability. How these interactive effects scale through different levels of ecological organization is poorly understood but essential for robust predictions of the effects of environmental change. In 2007, a year of record low rainfall, we conducted a prescribed fire in a coastal grassland in Orange County California. Within both burned and unburned portions of the grassland, we increased and decreased rainfall (with water addition and rainout shelters, respectively) and increased nitrogen (with N fertilization) in all possible treatment combinations. We asked the question: can physiological responses of the dominant species predict changes in ecosystem function to these interactive environmental manipulations? The native perennial bunchgrass, Nassella pulchra had higher rates of CO 2 uptake and stomatal conductance than the nonnative annual grass, Bromus diandrus across rainfall treatments in both the burned and unburned areas. Both species maintained relatively constant physiological responses regardless of environmental manipulation. Thus, based on these resilient individual-level responses, we predicted that ecosystem-responses would be relatively resilient to the environmental changes. Consistent with this prediction, burning and nitrogen did not strongly affect ecosystem function. However, we detected relatively large responses at the ecosystem level in response to rainfall manipulations, and these effects were generally consistent across burning and N fertilization treatments. Ecosystem respiration, photosynthesis (GEE) and evapotranspiration (ET) declined in response to rainfall removal but did not respond to increases in rainfall. In contrast, the response of annual net primary productivity (ANPP) and a canopy spectral index (NDVI) was greatest in rainfall addition plots. NDVI was correlated with ANPP, GEE and ET in both experiments. During this extremely dry growing season, many on the ecosystem responses were likely driven by soil microbial and plant population- level changes rather than changes in individual plant physiology. Ongoing research at the site will determine whether these relationships will change with annual climatic variation and as longer-term processes (e.g., species turnover, organic matter) begin to contribute more to individual and ecosystem responses.
B43B-1171
Complex Interactions Among Resources Drive Grassland Ecosystem Responses in two Transplanting Experiments
Changes in patterns of above- and belowground biomass, biodiversity, plant species and guild composition, and biogeochemical cycles were assessed in two transplanting experiments in mesic mountain grasslands in the Pyrenees. In both experiments, turf sods were transplanted from upland to lowland locations. The first experiment aimed to evaluate general responses to warming and drought, and the second to disentangle the effects of possible underlying mechanisms through resource manipulation by means of a nitrogen x phosphorus fertilisation experiment. Mesic grasslands showed strong shifts in plant diversity and composition after a short period of warming and drought, as a consequence of acute vulnerability of some dominant grasses, rare species losses, and aggregate and trigger effects of originally uncommon forb species. Environmental factors interacted in complex ways, producing changes in biomass distribution and guild proportions. Grasses dominated in the upland and at high resource levels while forbs dominated in the lowland and when water and nutrients decreased. The increased aboveground biomass in grassland sods transplanted to the lowland suggests that biomass production was more temperature-limited than water-limited. The enhancement effect found in the upland sods following phosphorus fertilization supports the hypothesis of a strong limitation arising from reduced nutrient availability, confirming the central role played by phosphorus in these grasslands. Nitrogen addition did not stimulate total biomass but affected guild composition. The counterintuitive effect of increased biomass with decreased water in the lowland was related to shifts in dominance from grasses to forbs, probably enabled by decreased nutrient availability under drought conditions.
B43B-1172
Foraminiferal Assemblage Zones of Oregon Salt-Marshes: Implications for Studies of Relative Sea-Level Change
Salt-marsh foraminifera can be useful tools for reconstructing Holocene sea-level change. The use of fossil assemblages as a relative sea-level indicator is underpinned by (1) the relation between characteristic modern foraminiferal assemblages and the tidal frame, (2) that these modern relations can be quantified, and (3) that modern assemblages are representative of past conditions recorded in Holocene salt-marsh sedimentary sequences. Using correspondence and transfer function analysis on large data sets, the former elevation (relative to the tidal frame) of a preserved salt-marsh deposit can be determined from foraminifera therein. This measure of paleomarsh-surface elevation, when combined with other lithostratigraphic data, can be used to infer the past position of local relative sea-level. We have collected sediment samples along elevational transects across the modern surface of seven tidal salt- marshes along coastal Oregon for foraminifera and environmental variables (elevation relative to the tidal frame, loss on ignition, grain size, pH, salinity) to determine the local and regional foraminiferal assemblage zones. Dominant agglutinated taxa in salt-marshes include Balticammina pseudomacresens, Haplophragmoides wilberti, Trochammina inflata, and Miliammina fusca. We will use canonical correspondence analysis of the modern assemblage and environmental data to determine what variable(s) control foraminiferal distribution. Seasonal and temporal changes in assemblage and species infaunality can modify the assemblage zonation decreasing the precision of environmental reconstruction. Therefore we sampled our sites twice throughout the year and compared the collected faunal assemblages and assessed species infaunal migration through staining live foraminifera in 30cm short cores.
B43B-1173
Effects of invasive species on ecosystem carbon dynamics in a restored tallgrass prairie
Land cover is an important determinant of soil C storage and dynamics. Restoration of degraded ecosystems and soils represents a target sink for offsetting rising atmospheric CO2 levels by increasing carbon sequestration in soils. The Conservation Reserve Program (CRP) and other initiatives to halt land degradation after cessation of cultivation present opportunities to assess the C sequestration potential of restoration practices. Our aim is to study what key ecosystem and climatic components exert the largest leverage for these lands to be sustainable C sinks. When considering controls on ecosystem C cycling, biodiversity has the potential to be a strong biotic influence. Invasive species can disrupt ecosystem processes by exhibiting functional characteristics which are distinct from their native counterparts. Invasive species, while affecting nearly all ecosystems, may pose a particular threat to restorations and impact rates of C accrual. We measured net ecosystem production (NEP) at a 18 years-old restored tallgrass prairie using the eddy covariance technique coupled to biometric estimates of biomass and soil C in a two year study where climatic conditions and plant species dominance varied. In 2005, the prairie restoration was a strong C sink with a NEP 438 gCm-2, despite a pronounced spring drought. In 2006, with above normal precipitation, a Melilotus alba dominance dramatically reduced NEP when compared to 2005. The loss of ecosystem functional diversity that resulted from the dominance of the invasive M. alba led to a 42% reduction in the length of the photosynthetically active season, as compared to the previous year. These results suggest that understudied biotic limitations to NEP may outweigh the effects of more commonly studied abiotic limitations. Ecosystem models and management strategies should consider biotic limitations to NEP in grasslands in order to maximize long term C sequestration of restorations and CRP management practices.
B43B-1174
Mineralization of Soil Organic Matter in Two Elevated CO2 by Warming Experiments in Grassland
Experimentally elevated atmospheric CO2 has enhanced carbon (C) allocation belowground, while ecosystem warming has led to losses of soil C due to enhanced mineralization of soil organic matter (SOM). Few investigations of possible interactions between elevated CO2 and temperature have been reported, but the potential for C cycling effects not to be simply additive is high. We have taken advantage of two multi-factor global change experiments being conducted in mixed C3/C4 grasslands to evaluate similarities and differences in responses of SOM mineralization rates. The TasFACE experiment in Tasmania, Australia, has been running for over 5 years, while the Prairie Heating and CO2 Enrichment (PHACE) experiment in Wyoming, USA, has been running for less than 2 years. Both experiments employ mini-FACE systems (enriched plots targeted at 550 at TasFACE and 600 ppm at PHACE) and overhead ceramic infrared emitters (heated plots targeted at +2 degrees C at TasFACE and +1.5/+3 degrees day/night at PHACE). Soil samples collected after 5 years at TasFACE and at the beginning of the second year at PHACE were incubated for three weeks to evaluate changes in labile SOM pool sizes and turnover rates. We hypothesized that elevated CO2 would enhance labile SOM pool size and that warming would reduce it, and that warming would stimulate decomposition rate. Preliminary results suggested that five years of warming enhanced decomposition rate in the TasFACE soils, but only under the C4 grass species, whereas the first two months of warming had no effects on decomposition rate at PHACE. Elevated CO2 increased mineralizable C pool sizes by 10 to 30 percent, depending on depth, in the TasFACE soils, but did not significantly alter C cycling in the PHACE soils. Short experimental duration likely explained the lack of treatment effects seen at PHACE. We plan to continue conducting parallel experiments to track temporal changes in C cycling with the expectation that interactive effects of elevated CO2 and warming may appear over the long term.
B43B-1175 INVITED
Ecosystem Responses in the Jasper Ridge Global Change Experiment Over Seven Years
Field plots within a California grassland at Jasper Ridge, USA, have been exposed for 9 years to enhanced levels of CO2, nitrogen, heat, and precipitation, singly and in all possible combinations. We here report an updated analysis of major trends in NPP response, based on an analysis of data collected through 2006. First, as reported in earlier syntheses of NPP response at the JRGCE, NPP responded most strongly and consistently to enhanced N deposition. However, the magnitude of this response peaked in 2003 and had been markedly less pronounced since then. Second, all statistically significant effects emerged as interactions with year, highlighting the dependence of ecosystem responses to global change factors on temporal variation in uncontrolled drivers. Third, the influence of enhanced CO2 on NPP varied regularly across years, and correlated strongly with annual precipitation (R2=0.89, p<0.05). Contrary to expectations, enriched CO2 concentrations diminished NPP in dry years and enhanced NPP in wet years. Two observations point to the seasonal pattern of precipitation (vs. total amount) as mediating the ecosystem response to interactions among global change factors. First, although the effect of CO2 varied regularly with annual precipitation, the interaction of the CO2 and precipitation treatments did not produce a parallel result. The enhanced precipitation treatment in the JRGCE adds 50% to each rain event and adds two rain events to the end of the growing season but does not shift the pattern of rainfall during the growing season. Second, enhanced CO2 appears to diminish the positive effect of N deposition on NPP, contrary to a traditional model of co-limitation by C and N, but only in years with minimal precipitation during March. http://globalecology.stanford.edu/DGE/Dukes/JRGCE/home.html