Quantifying shrub-shrub competition in drylands using aerial imagery and a novel landscape competition index.

The Jornada Basin Long-Term Ecological Research Site (JRN-LTER, or JRN) is a semiarid grassland-shrubland in southern New Mexico, USA. The role of intraspecific competition in constraining shrub growth and establishment at the JRN and in arid systems, in general, is an important question in dryland studies. Using information on shrub distributions and growth habits at the JRN, we present a novel landscape-scale (c. 1 ha) metric (the 'competition index', CI), which quantifies the potential intensity of competitive interactions. We map and compare the intensity of honey mesquite (Prosopis glandulosa, Torr.) competition spatially and temporally across the JRN-LTER, investigating associations of CI with shrub distribution, density, and soil types. The CI metric shows strong correlation with values of percent cover. Mapping CI across the Jornada Basin shows that high-intensity intraspecific competition is not prevalent, with few locations where intense competition is likely to be limiting further honey mesquite expansion. Comparison of CI among physiographic provinces shows differences in average CI values associated with geomorphology, topography, and soil type, suggesting that edaphic conditions may impose important constraints on honey mesquite and growth. However, declining and negative growth rates with increasing CI suggest that intraspecific competition constrains growth rates when CI increases above c. 0.5.

Global application of an unoccupied aerial vehicle photogrammetry protocol for predicting aboveground biomass in non-forest ecosystems.

Non-forest ecosystems, dominated by shrubs, grasses and herbaceous plants, provide ecosystem services including carbon sequestration and forage for grazing, and are highly sensitive to climatic changes. Yet these ecosystems are poorly represented in remotely sensed biomass products and are undersampled by in situ monitoring. Current global change threats emphasize the need for new tools to capture biomass change in non-forest ecosystems at appropriate scales. Here we developed and deployed a new protocol for photogrammetric height using unoccupied aerial vehicle (UAV) images to test its capability for delivering standardized measurements of biomass across a globally distributed field experiment. We assessed whether canopy height inferred from UAV photogrammetry allows the prediction of aboveground biomass (AGB) across low-stature plant species by conducting 38 photogrammetric surveys over 741 harvested plots to sample 50 species. We found mean canopy height was strongly predictive of AGB across species, with a median adjusted R 2 of 0.87 (ranging from 0.46 to 0.99) and median prediction error from leave-one-out cross-validation of 3.9%. Biomass per-unit-of-height was similar within but different among, plant functional types. We found that photogrammetric reconstructions of canopy height were sensitive to wind speed but not sun elevation during surveys. We demonstrated that our photogrammetric approach produced generalizable measurements across growth forms and environmental settings and yielded accuracies as good as those obtained from in situ approaches. We demonstrate that using a standardized approach for UAV photogrammetry can deliver accurate AGB estimates across a wide range of dynamic and heterogeneous ecosystems. Many academic and land management institutions have the technical capacity to deploy these approaches over extents of 1-10 ha-1. Photogrammetric approaches could provide much-needed information required to calibrate and validate the vegetation models and satellite-derived biomass products that are essential to understand vulnerable and understudied non-forested ecosystems around the globe.

Study becomes insight: Ecological learning from machine learning.

The ecological and environmental science communities have embraced machine learning (ML) for empirical modelling and prediction. However, going beyond prediction to draw insights into underlying functional relationships between response variables and environmental 'drivers' is less straightforward. Deriving ecological insights from fitted ML models requires techniques to extract the 'learning' hidden in the ML models.We revisit the theoretical background and effectiveness of four approaches for deriving insights from ML: ranking independent variable importance (Gini importance, GI; permutation importance, PI; split importance, SI; and conditional permutation importance, CPI), and two approaches for inference of bivariate functional relationships (partial dependence plots, PDP; and accumulated local effect plots, ALE). We also explore the use of a surrogate model for visualization and interpretation of complex multi-variate relationships between response variables and environmental drivers. We examine the challenges and opportunities for extracting ecological insights with these interpretation approaches. Specifically, we aim to improve interpretation of ML models by investigating how effectiveness relates to (a) interpretation algorithm, (b) sample size and (c) the presence of spurious explanatory variables.We base the analysis on simulations with known underlying functional relationships between response and predictor variables, with added white noise and the presence of correlated but non-influential variables. The results indicate that deriving ecological insight is strongly affected by interpretation algorithm and spurious variables, and moderately impacted by sample size. Removing spurious variables improves interpretation of ML models. Meanwhile, increasing sample size has limited value in the presence of spurious variables, but increasing sample size does improves performance once spurious variables are omitted. Among the four ranking methods, SI is slightly more effective than the other methods in the presence of spurious variables, while GI and SI yield higher accuracy when spurious variables are removed. PDP is more effective in retrieving underlying functional relationships than ALE, but its reliability declines sharply in the presence of spurious variables. Visualization and interpretation of the interactive effects of predictors and the response variable can be enhanced using surrogate models, including three-dimensional visualizations and use of loess planes to represent independent variable effects and interactions.Machine learning analysts should be aware that including correlated independent variables in ML models with no clear causal relationship to response variables can interfere with ecological inference. When ecological inference is important, ML models should be constructed with independent variables that have clear causal effects on response variables. While interpreting ML models for ecological inference remains challenging, we show that careful choice of interpretation methods, exclusion of spurious variables and adequate sample size can provide more and better opportunities to 'learn from machine learning'.

Woody-biomass projections and drivers of change in sub-Saharan Africa.

Africa's ecosystems have an important role in global carbon dynamics, yet consensus is lacking regarding the amount of carbon stored in woody vegetation and the potential impacts to carbon storage in response to changes in climate, land use, and other Anthropocene risks. Here, we explore the socio-environmental conditions that shaped the contemporary distribution of woody vegetation across sub-Saharan Africa and evaluate ecosystem response to multiple scenarios of climate change, anthropogenic pressures, and fire disturbance. Our projections suggest climate change will have a small but negative effect on above ground woody biomass at the continental scale, and the compounding effects of population growth, increasing human pressures, and socio-climatic driven changes in fire behavior further exacerbate climate-driven trends. Relatively modest continental-scale trends obscure much larger regional perturbations, with climatic and anthropogenic factors leading to increased carbon storage potential in East Africa, offset by large deficits in West, Central, and Southern Africa.

Author Correction: HYSOGs250m, global gridded hydrologic soil groups for curve-number-based runoff modeling.

The original version of this Data Descriptor incorrectly referenced the "United Nations (UN) Food and Agriculture Organization (FAO) soilGrids250m system". This has been corrected to "SoilGrids predictions" throughout the text in both the HTML and PDF versions.

Change in dominance determines herbivore effects on plant biodiversity.

Herbivores alter plant biodiversity (species richness) in many of the world's ecosystems, but the magnitude and the direction of herbivore effects on biodiversity vary widely within and among ecosystems. One current theory predicts that herbivores enhance plant biodiversity at high productivity but have the opposite effect at low productivity. Yet, empirical support for the importance of site productivity as a mediator of these herbivore impacts is equivocal. Here, we synthesize data from 252 large-herbivore exclusion studies, spanning a 20-fold range in site productivity, to test an alternative hypothesis-that herbivore-induced changes in the competitive environment determine the response of plant biodiversity to herbivory irrespective of productivity. Under this hypothesis, when herbivores reduce the abundance (biomass, cover) of dominant species (for example, because the dominant plant is palatable), additional resources become available to support new species, thereby increasing biodiversity. By contrast, if herbivores promote high dominance by increasing the abundance of herbivory-resistant, unpalatable species, then resource availability for other species decreases reducing biodiversity. We show that herbivore-induced change in dominance, independent of site productivity or precipitation (a proxy for productivity), is the best predictor of herbivore effects on biodiversity in grassland and savannah sites. Given that most herbaceous ecosystems are dominated by one or a few species, altering the competitive environment via herbivores or by other means may be an effective strategy for conserving biodiversity in grasslands and savannahs globally.

HYSOGs250m, global gridded hydrologic soil groups for curve-number-based runoff modeling.

Hydrologic soil groups (HSGs) are a fundamental component of the USDA curve-number (CN) method for estimation of rainfall runoff; yet these data are not readily available in a format or spatial-resolution suitable for regional- and global-scale modeling applications. We developed a globally consistent, gridded dataset defining HSGs from soil texture, bedrock depth, and groundwater. The resulting data product-HYSOGs250m-represents runoff potential at 250 m spatial resolution. Our analysis indicates that the global distribution of soil is dominated by moderately high runoff potential, followed by moderately low, high, and low runoff potential. Low runoff potential, sandy soils are found primarily in parts of the Sahara and Arabian Deserts. High runoff potential soils occur predominantly within tropical and sub-tropical regions. No clear pattern could be discerned for moderately low runoff potential soils, as they occur in arid and humid environments and at both high and low elevations. Potential applications of this data include CN-based runoff modeling, flood risk assessment, and as a covariate for biogeographical analysis of vegetation distributions.

Comment on "The extent of forest in dryland biomes".

Bastin et al (Reports, 12 May 2017, p. 635) infer forest as more globally extensive than previously estimated using tree cover data. However, their forest definition does not reflect ecosystem function or biotic composition. These structural and climatic definitions inflate forest estimates across the tropics and undermine conservation goals, leading to inappropriate management policies and practices in tropical grassy ecosystems.

Spatial vegetation patterns and neighborhood competition among woody plants in an East African savanna.

The majority of research on savanna vegetation dynamics has focused on the coexistence of woody and herbaceous vegetation. Interactions among woody plants in savannas are relatively poorly understood. We present data from a 10-yr longitudinal study of spatially explicit growth patterns of woody vegetation in an East African savanna following exclusion of large herbivores and in the absence of fire. We examined plant spatial patterns and quantified the degree of competition among woody individuals. Woody plants in this semiarid savanna exhibit strongly clumped spatial distributions at scales of 1-5 m. However, analysis of woody plant growth rates relative to their conspecific and heterospecific neighbors revealed evidence for strong competitive interactions at neighborhood scales of up to 5 m for most woody plant species. Thus, woody plants were aggregated in clumps despite significantly decreased growth rates in close proximity to neighbors, indicating that the spatial distribution of woody plants in this region depends on dispersal and establishment processes rather than on competitive, density-dependent mortality. However, our documentation of suppressive effects of woody plants on neighbors also suggests a potentially important role for tree-tree competition in controlling vegetation structure and indicates that the balanced-competition hypothesis may contribute to well-known patterns in maximum tree cover across rainfall gradients in Africa.

On regreening and degradation in Sahelian watersheds.

Over many decades our understanding of the impacts of intermittent drought in water-limited environments like the West African Sahel has been influenced by a narrative of overgrazing and human-induced desertification. The desertification narrative has persisted in both scientific and popular conception, such that recent regional-scale recovery ("regreening") and local success stories (community-led conservation efforts) in the Sahel, following the severe droughts of the 1970s-1980s, are sometimes ignored. Here we report a study of watershed-scale vegetation dynamics in 260 watersheds, sampled in four regions of Senegal, Mali, and Niger from 1983-2012, using satellite-derived vegetation indices as a proxy for net primary production. In response to earlier controversy, we first examine the shape of the rainfall-net primary production relationship and how it impacts conclusions regarding greening or degradation. We conclude that the choice of functional relationship has little quantitative impact on our ability to infer greening or degradation trends. We then present an approach to analyze changes in long-term (decade-scale) average rain-use efficiency (an indicator of slowly responding vegetation structural changes) relative to changes in interannual-scale rainfall sensitivity (an indicator of landscape ability to respond rapidly to rainfall variability) to infer trends in greening/degradation of the watersheds in our sample regions. The predominance of increasing rain-use efficiency in our data supports earlier reports of a "greening" trend across the Sahel. However, there are strong regional differences in the extent and direction of change, and in the apparent role of changing woody and herbaceous components in driving those temporal trends.

Effects of tree harvest on the stable-state dynamics of savanna and forest.

Contemporary theory on the maintenance and stability of the savanna biome has focused extensively on how climate and disturbances interact to affect tree growth and demography. In particular, the role of fire in reducing tree cover from climatic maxima is now well appreciated, and in certain cases, herbivory also strongly affects tree cover. However, in African savannas and forests, harvest of trees by humans for cooking and heating is an oft overlooked disturbance. Thus, we incorporate tree harvest into a population dynamic model of grasses, savanna saplings, savanna trees, and forest trees. We use assumptions about the differential demographic responses of savanna trees and forest trees to harvest to show how tree harvest influences tree cover, demography, and community composition. Tree harvest can erode the intrinsic basin of attraction for forest and make a state transition via fire to savanna more likely. The savanna state is generally resilient to all but high levels of tree harvest because of the resprouting abilities of savanna trees. In the absence of active fire suppression, our analysis suggests that we can expect to see large and potentially irreversible shifts from forest to savanna as demand increases for charcoal in sub-Saharan Africa. On the other hand, savanna tree species' traits promote savanna stability in the face of low to moderate harvest pressure.

Savanna vegetation-fire-climate relationships differ among continents.

Ecologists have long sought to understand the factors controlling the structure of savanna vegetation. Using data from 2154 sites in savannas across Africa, Australia, and South America, we found that increasing moisture availability drives increases in fire and tree basal area, whereas fire reduces tree basal area. However, among continents, the magnitude of these effects varied substantially, so that a single model cannot adequately represent savanna woody biomass across these regions. Historical and environmental differences drive the regional variation in the functional relationships between woody vegetation, fire, and climate. These same differences will determine the regional responses of vegetation to future climates, with implications for global carbon stocks.

Analysis of stable states in global savannas: is the CART pulling the horse?

Multiple stable states, bifurcations and thresholds are fashionable concepts in the ecological literature, a recognition that complex ecosystems may at times exhibit the interesting dynamic behaviours predicted by relatively simple biomathematical models. Recently, several papers in Global Ecology and Biogeography, Proceedings of the National Academy of Sciences USA, Science and elsewhere have attempted to quantify the prevalence of alternate stable states in the savannas of Africa, Australia and South America, and the tundra-taiga-grassland transitions of the circum-boreal region using satellite-derived woody canopy cover. While we agree with the logic that basins of attraction can be inferred from the relative frequencies of ecosystem states observed in space and time, we caution that the statistical methodologies underlying the satellite product used in these studies may confound our ability to infer the presence of multiple stable states. We demonstrate this point using a uniformly distributed 'pseudo-tree cover' database for Africa that we use to retrace the steps involved in creation of the satellite tree-cover product and subsequent analysis. We show how classification and regression tree (CART)-based products may impose discontinuities in satellite tree-cover estimates even when such discontinuities are not present in reality. As regional and global remote sensing and geospatial data become more easily accessible for ecological studies, we recommend careful consideration of how error distributions in remote sensing products may interact with the data needs and theoretical expectations of the ecological process under study.

Allometric convergence in savanna trees and implications for the use of plant scaling models in variable ecosystems.

Theoretical models of allometric scaling provide frameworks for understanding and predicting how and why the morphology and function of organisms vary with scale. It remains unclear, however, if the predictions of 'universal' scaling models for vascular plants hold across diverse species in variable environments. Phenomena such as competition and disturbance may drive allometric scaling relationships away from theoretical predictions based on an optimized tree. Here, we use a hierarchical Bayesian approach to calculate tree-specific, species-specific, and 'global' (i.e. interspecific) scaling exponents for several allometric relationships using tree- and branch-level data harvested from three savanna sites across a rainfall gradient in Mali, West Africa. We use these exponents to provide a rigorous test of three plant scaling models (Metabolic Scaling Theory (MST), Geometric Similarity, and Stress Similarity) in savanna systems. For the allometric relationships we evaluated (diameter vs. length, aboveground mass, stem mass, and leaf mass) the empirically calculated exponents broadly overlapped among species from diverse environments, except for the scaling exponents for length, which increased with tree cover and density. When we compare empirical scaling exponents to the theoretical predictions from the three models we find MST predictions are most consistent with our observed allometries. In those situations where observations are inconsistent with MST we find that departure from theory corresponds with expected tradeoffs related to disturbance and competitive interactions. We hypothesize savanna trees have greater length-scaling exponents than predicted by MST due to an evolutionary tradeoff between fire escape and optimization of mechanical stability and internal resource transport. Future research on the drivers of systematic allometric variation could reconcile the differences between observed scaling relationships in variable ecosystems and those predicted by ideal models such as MST.

Nutrient resorption patterns of plant functional groups in a tropical savanna: variation and functional significance.

Green and senesced leaf nitrogen (N) and phosphorus (P) concentrations of different plant functional groups in savanna communities of Kruger National Park, South Africa were analyzed to determine if nutrient resorption was regulated by plant nutritional status and foliar N:P ratios. The N and P concentrations in green leaves and the N concentrations in senesced leaves differed significantly between the dominant plant functional groups in these savannas: fine-leaved trees, broad-leaved trees and grasses. However, all three functional groups reduced P to comparable and very low levels in senesced leaves, suggesting that P was tightly conserved in this tropical semi-arid savanna ecosystem. Across all functional groups, there was evidence for nutritional control of resorption in this system, with both N and P resorption efficiencies decreasing as green leaf nutrient concentrations increased. However, specific patterns of resorption and the functional relationships between nutrient concentrations in green and senesced leaves varied by nutrient and plant functional group. Functional relationships between N concentrations in green and senesced leaves were indistinguishable between the dominant groups, suggesting that variation in N resorption efficiency was largely the result of inter-life form differences in green leaf N concentrations. In contrast, observed differences in P resorption efficiencies between life forms appear to be the result of both differences in green leaf P concentrations as well as inherent differences between life forms in the fraction of green leaf P resorbed from senescing leaves. Our results indicate that foliar N:P ratios are poor predictors of resorption efficiency in this ecosystem, in contrast to N and P resorption proficiencies, which are more responsive to foliar N:P ratios.

Do fires in savannas consume woody biomass? A comment on approaches to modeling savanna dynamics.

Savanna ecosystems have long been fertile ground for mathematical modeling of vegetation structure and the role of resources and disturbance in tree-grass coexistence. In recent years, several authors have presented models that explore how savanna fires suppress the woody community, alter ecosystem dynamics, and promote grass persistence. We argue, however, that the assumption that fires influence savanna dynamics by consuming woody biomass may be wrong because, in reality, fires kill seedlings and saplings that constitute little biomass relative to adult trees. We present a simple alternative that separates the woody community into a subadult (fire-sensitive) class and an adult (fire-resistant) class and explore how this ecologically more realistic, but still simplified, model may provide better simulations of demographic processes and response to fires in savannas.