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

In their recent paper, Hanan et al. (Global Ecology and Biogeography, 2014, 23 , 259–263) argue that the use of classification and regression trees (CARTs) to calibrate global remote sensing datasets, including the MODIS VCF tree‐cover dataset, makes these data inappropriate for analysing the frequency distribution of tree cover. While we agree with their most general point – that the use of remote sensing products should be informed and deliberate – their analysis overlooks a few key aspects of the use of CARTs in generating global tree‐cover data. Firstly, while their presentation of flaws in the use of CARTs is compelling, their use of hypothetical data obscures the reasons why CARTs are a useful tool. Secondly, they do not actually examine the error distributions of the MODIS VCF tree‐cover data. Such an analysis, which we perform, revealed the following: (1) the MODIS VCF product may not be useful for differentiating over small ranges of tree cover (less than c. 10%); (2) that the bimodality of low and high tree cover, with a frequency minimum at intermediate tree cover, is not attributable to bias in MODIS VCF tree‐cover calibrations; and (3) that the MODIS VCF is not well‐resolved below c. 20–30% tree cover, such that MODIS cannot be used with any confidence to evaluate multimodality in tree cover in that range. Further validation and calibration are likely to be helpful and, at low tree cover, necessary for improving MODIS VCF tree‐cover estimates. However, the MODIS VCF – which has facilitated major steps in our ability to examine ecological phenomena at global scales – remains a useful tool for well‐informed ecological analysis.     

Analysis of stable states in global savannas – a response to Staver and Hansen

Staver & Hansen (2015, Global Ecology and Biogeography, doi: 10.1111/geb.12285) comment on our recent paper (Hanan et al., Global Ecology and Biogeography, 2014, 23 , 259–263) in which we argue that classification and regression tree methods used with remote sensing data to predict tree cover may bias inference of bifurcations in savanna vegetation communities. While we agree with several of their comments, we remain unconvinced that a remote sensing product based on an inherently discontinuous statistical approach can, or should, be used to test for discontinuities.     

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.     

Large‐scale patterns of tree species richness and the metabolic theory of ecology

The metabolic theory of ecology (MTE) endeavours to explain ecosystem structure and function in terms of the effects of temperature and body size on metabolic rate. In a recent paper (Wang et al., 2009, Proceedings of the National Academy of Sciences USA, 106 , 13388), we tested the MTE predictions of species richness using tree distributions in eastern Asia and North America. Our results supported the linear relationship between log‐transformed species richness and the inverse of absolute temperature predicted by the MTE, but the slope strongly depends on spatial scale. The results also indicate that there are more tree species in cold climate at high latitudes in North America than in eastern Asia, but the reverse is true in warm climate at low latitudes. Qian & Ricklefs (2011, Global Ecology and Biogeography, 20 , 362–365) recently questioned our data and some of the analyses. Here we reply to them, and provide further analyses to show that their critiques are primarily based on unsuitable data and subjective conjecture.     

A Changing Number of Alternative States in the Boreal Biome: Reproducibility Risks of Replacing Remote Sensing Products

Publicly available remote sensing products have boosted science in many ways. The openness of these data sources suggests high reproducibility. However, as we show here, results may be specific to versions of the data products that can become unavailable as new versions are posted. We focus on remotely-sensed tree cover. Recent studies have used this public resource to detect multi-modality in tree cover in the tropical and boreal biomes. Such patterns suggest alternative stable states separated by critical tipping points. This has important implications for the potential response of these ecosystems to global climate change. For the boreal region, four distinct ecosystem states (i.e., treeless, sparse and dense woodland, and boreal forest) were previously identified by using the Collection 3 data of MODIS Vegetation Continuous Fields (VCF). Since then, the MODIS VCF product has been updated to Collection 5; and a Landsat VCF product of global tree cover at a fine spatial resolution of 30 meters has been developed. Here we compare these different remote-sensing products of tree cover to show that identification of alternative stable states in the boreal biome partly depends on the data source used. The updated MODIS data and the newer Landsat data consistently demonstrate three distinct modes around similar tree-cover values. Our analysis suggests that the boreal region has three modes: one sparsely vegetated state (treeless), one distinct ‘savanna-like’ state and one forest state, which could be alternative stable states. Our analysis illustrates that qualitative outcomes of studies may change fundamentally as new versions of remote sensing products are used. Scientific reproducibility thus requires that old versions remain publicly available.     

High‐severity fire corroborated in historical dry forests of the western United States: response to Fulé et al.

Accurate assessment of changing fire regimes is important, since climatic change and people may be promoting more wildfires. Government wildland fire policies and restoration programmes in dry western US forests are based on the hypothesis that high‐severity fire was rare in historical fire regimes, modern fire severity is unnaturally high and restoration efforts should focus primarily on thinning forests to eliminate high‐severity fire. Using General Land Office (GLO) survey data over large dry‐forest landscapes, we showed that the proportion of historical forest affected by high‐severity fire was not insignificant, fire severity has not increased as a proportion of total fire area and large areas of dense forest were present historically (Williams & Baker, Global Ecology and Biogeography, 21 , 1042–1052, 2012; W&B). In response, Fulé et al. (Global Ecology and Biogeography, 2013, doi: 10.1111/geb.12136; FE) suggest that our inferences are unsupported and land management based on our research could be damaging to native ecosystems. Here, we show that the concerns of FE are unfounded. Their criticism comes from misquoting W&B, mistaking W&B's methods, misusing evidence (e.g. from Aldo Leopold) and missing substantial available evidence. We also update corroboration for the extensive historical high‐severity fire shown by W&B. We suggest that restoration programmes are misdirected in seeking to reduce all high‐severity fire in dry forests, given findings from spatially extensive GLO data and other sources.     

Response to commentary by Woinarski (Critical‐weight‐range marsupials in northern Australia are declining: a commentary on Fisher et al. (2014) ‘The current decline of tropical marsupials in Australia: is history repeating?’)

The recent commentary by Woinarski (2014, Global Ecology and Biogeography, doi: 10.1111/geb.12165) disagreed with our conclusions on the correlates of decline in the marsupials of tropical Australia (Fisher et al., 2014, Global Ecology and Biogeography, 23 , 181–190). We compared traits of species that were associated with range decline in southern and northern Australia. We found that habitat structure, climate and body size were correlated with range decline. In the north, declines of marsupials were most severe in savanna with moderate rainfall. In the south, the ranges of species in open habitat with very low rainfall have declined most. Also, the association between range decline and body mass differed between north and south: this is the main concern of Woinarski, who further disagreed with our choice of the Tropic of Capricorn as a boundary between north and south, our omission of rodents, how to treat timing of extinctions, and our inference that cats are major drivers of decline. We address these concerns in this response.     

Unsupported inferences of high‐severity fire in historical dry forests of the western United States: response to Williams and Baker

Reconstructions of dry western US forests in the late 19th century in Arizona, Colorado and Oregon based on General Land Office records were used by Williams & Baker (2012; Global Ecology and Biogeography, 21 , 1042–1052; hereafter W&B) to infer past fire regimes with substantial moderate and high‐severity burning. The authors concluded that present‐day large, high‐severity fires are not distinguishable from historical patterns. We present evidence of important errors in their study. First, the use of tree size distributions to reconstruct past fire severity and extent is not supported by empirical age–size relationships nor by studies that directly quantified disturbance history in these forests. Second, the fire severity classification of W&B is qualitatively different from most modern classification schemes, and is based on different types of data, leading to an inappropriate comparison. Third, we note that while W&B asserted ‘surprising’ heterogeneity in their reconstructions of stand density and species composition, their data are not substantially different from many previous studies which reached very different conclusions about subsequent forest and fire behaviour changes. Contrary to the conclusions of W&B, the preponderance of scientific evidence indicates that conservation of dry forest ecosystems in the western United States and their ecological, social and economic value is not consistent with a present‐day disturbance regime of large, high‐severity fires, especially under changing climate.     

Benthic habitats do show a significant latitudinal diversity gradient: A comment on Kinlock et al. (2018)

The latitudinal diversity gradient (LDG) has been investigated for decades, with hundreds of studies focusing on different organisms, regions and habitat types. Meta‐analysis may be considered, therefore, a useful tool to explore the generality and limitations of this remarkable macroecological pattern. The first meta‐analysis exploring variations in the LDG, published by Hillebrand in 2004, revealed that the latitudinal decline in species richness seems to be indeed a general phenomenon. However, Kinlock et al. (2018, Global Ecology and Biogeography, 27, 125–141) revisited recently the challenge of synthesizing individual LDGs and indicated that the phenomenon is not ubiquitous among habitats of the marine realm. More precisely, they indicated that the phenomenon is non‐significant in the benthic habitat. Here, we suggest that the marine habitat categories used by them (i.e., benthic, coral reefs, coastal, open ocean) are not independent and that reclassifying the studies significantly alters one of their main results. By assigning the studies into benthic and pelagic categories, and additionally into coastal or oceanic zones, we show that non‐ambiguous, evolutionarily meaningful marine habitats display a significant latitudinal decline in species richness.     

Rethinking the relationship between nestedness and beta diversity: a comment on Baselga (2010)

Baselga [Partitioning the turnover and nestedness components of beta diversity. Global Ecology and Biogeography, 19 , 134–143, 2010] proposed pairwise (β_nes) and multiple‐site (β_NES) beta‐diversity measures to account for the nestedness component of beta diversity. We used empirical, randomly created and idealized matrices to show that both measures are only partially related to nestedness and do not fit certain fundamental requirements for consideration as true nestedness‐resultant dissimilarity measures. Both β_nes and β_NES are influenced by matrix size and fill, and increase or decrease even when nestedness remains constant. Additionally, we demonstrate that β_NES can yield high values even for matrices with no nestedness. We conclude that β_nes and β_NES are not true measures of the nestedness‐resultant dissimilarity between sites. Actually, they quantify how differences in species richness that are not due to species replacement contribute to patterns of beta diversity. Finally, because nestedness is a special case of dissimilarity in species composition due to ordered species loss (or gain), the extent to which differences in species composition is due to nestedness can be measured through an index of nestedness.     

The need for richness‐independent measures of turnover when delineating biogeographical regions

Delineating biogeographical regions is one of the primary steps when analysing biogeographical patterns. In their proposed quantitative framework, Kreft & Jetz (2010, Journal of Biogeography, 37 , 2029–2053) recommended the use of the β_sim index to delineate biogeographical regions because this turnover measure is weakly affected by differences in species richness between localities. A recent study by Carvalho et al. (2012, Global Ecology and Biogeography, 21 , 760–771) critiziced the use of β_sim in ecological and biogeographical studies, and proposed the β_‐3 index. Here we used simple numerical examples and an empirical case study (European freshwater fishes) to highlight potential pitfalls associated with the use of β_‐3 for bioregionalization. We show that β_‐3 is not a richness‐independent measure of species turnover. We also show that this index violates the ‘complementarity’ property, namely that localities without species in common have the largest dissimilarity, which is an essential prerequisite for beta diversity studies.     

Critical‐weight‐range marsupials in northern Australia are declining: a commentary on Fisher et al. (2014) ‘The current decline of tropical marsupials in Australia: is history repeating?’

Many mammal species are declining in parts of Australia's tropical savannas, for reasons that are not yet well defined. A recent paper (Fisher et al., 2014, Global Ecology and Biogeography, 23 , 181–190) suggested that the primary cause is predation by feral cats, with the main evidence presented being a purported over‐representation of small species amongst the marsupials that have contracted in range (‘small body size signifies high current extinction risk’). However, a review here of the information presented in that paper shows that no marsupial species smaller than 100 g has shown range contraction in northern Australia, and that most (15 of 17) declines are of species in the ‘critical weight range’ (35 g to 5.5 kg).     

Classification and regression tree (CART) analyses of genomic signatures reveal sets of tetramers that discriminate temperature optima of archaea and bacteria

Classification and regression tree (CART) analysis was applied to genome-wide tetranucleotide frequencies (genomic signatures) of 195 archaea and bacteria. Although genomic signatures have typically been used to classify evolutionary divergence, in this study, convergent evolution was the focus. Temperature optima for most of the organisms examined could be distinguished by CART analyses of tetranucleotide frequencies. This suggests that pervasive (nonlinear) qualities of genomes may reflect certain environmental conditions (such as temperature) in which those genomes evolved. The predominant use of GAGA and AGGA as the discriminating tetramers in CART models suggests that purine-loading and codon biases of thermophiles may explain some of the results.     

Characterizing abundance–occupancy relationships: there is no artefact

Positive abundance–occupancy relationships (AORs) are among the most general macroecological patterns: locally common species are regionally widespread, locally rare species are regionally restricted. In a recent contribution, Wilson (Global Ecology and Biogeography, 2011, 20 , 193–202) made three claims: (1) that AORs are critically dependent on the method used to calculate average abundance; (2) averaging abundance over occupied sites tends to lead to a very high incidence of negative relationships; (3) this represents a statistical artefact that should be considered in studies of AORs. Here we show that this outcome arises in Wilson's simulations purely due to an arbitrary choice of occupancy models and parameter ranges. The resulting negative relationships are not statistical artefacts, but are easily interpreted in terms of spatial aggregation in abundant species. The fact that empirical evidence fails to support a high prevalence of negative AORs suggests, however, that such parameter combinations arise only rarely in nature. We conclude that simulations that are based on untested assumptions, and that produce patterns unsupported by empirical evidence, have limited use in characterizing AORs, and add little to understanding of the processes driving important relationships between local population size and regional occupancy.     

Panbiogeography of New Caledonia: a response to Heads (2008)

The definition of areas of endemism is one of the most important steps for historical biogeography. Here I review the dataset used by Heads (Journal of Biogeography, 2008, 35 , 2153–2175) for his panbiogeographical analysis of the New Caledonian biota. I highlight that some of the distribution data appear dubious (some localities should have been included while some others should have been deleted) when compared with current databases. In addition, some conclusions are not supported by the data themselves.     

Tree demography dominates long‐term growth trends inferred from tree rings

Understanding responses of forests to increasing CO₂ and temperature is an important challenge, but no easy task. Tree rings are increasingly used to study such responses. In a recent study, van der Sleen et al. (2014) Nature Geoscience, 8, 4 used tree rings from 12 tropical tree species and find that despite increases in intrinsic water use efficiency, no growth stimulation is observed. This challenges the idea that increasing CO₂ would stimulate growth. Unfortunately, tree ring analysis can be plagued by biases, resulting in spurious growth trends. While their study evaluated several biases, it does not account for all. In particular, one bias may have seriously affected their results. Several of the species have recruitment patterns, which are not uniform, but clustered around one specific year. This results in spurious negative growth trends if growth rates are calculated in fixed size classes, as ‘fast‐growing’ trees reach the sampling diameter earlier compared to slow growers and thus fast growth rates tend to have earlier calendar dates. We assessed the effect of this ‘nonuniform age bias’ on observed growth trends and find that van der Sleen's conclusions of a lack of growth stimulation do not hold. Growth trends are – at least partially – driven by underlying recruitment or age distributions. Species with more clustered age distributions show more negative growth trends, and simulations to estimate the effect of species’ age distributions show growth trends close to those observed. Re‐evaluation of the growth data and correction for the bias result in significant positive growth trends of 1–2% per decade for the full period, and 3–7% since 1950. These observations, however, should be taken cautiously as multiple biases affect these trend estimates. In all, our results highlight that tree ring studies of long‐term growth trends can be strongly influenced by biases if demographic processes are not carefully accounted for.     

Response to Collins et al. (2011)

We note serious problems in Collins et al. (Journal of Biogeography, 2011, doi: 10.1111/j.1365‐2699.2011.02506.x ): failure to use over 80% of the available data; failure to use one of the two available archipelagoes; mistaken inclusion of four species; and reliance on a grossly inadequate number of null matrices. Curing the paper of these problems would have strengthened the evidence for checkerboards and the role of competition.     

Seeing the forest beyond the trees

In a recent paper (Mitchard et al. 2014, Global Ecology and Biogeography, 23 , 935–946) a new map of forest biomass based on a geostatistical model of field data for the Amazon (and surrounding forests) was presented and contrasted with two earlier maps based on remote‐sensing data Saatchi et al. (2011; RS1) and Baccini et al. (2012; RS2). Mitchard et al. concluded that both the earlier remote‐sensing based maps were incorrect because they did not conform to Mitchard et al. interpretation of the field‐based results. In making their case, however, they misrepresented the fundamental nature of primary field and remote‐sensing data and committed critical errors in their assumptions about the accuracy of research plots, the interpolation methodology and the statistical analysis. By ignoring the large uncertainty associated with ground estimates of biomass and the significant under‐sampling and spatial bias of research plots, Mitchard et al. reported erroneous trends and artificial patterns of biomass over Amazonia. Because of these misrepresentations and methodological flaws, we find their critique of the satellite‐derived maps to be invalid.