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

Abstract

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.

MODIS VCF should not be used to detect discontinuities in tree cover due to binning bias. A comment on Hanan et al. (2014) and Staver and Hansen (2015)

“…Discontinuities in tree cover frequency distributions derived from this data have been used to support the hypothesis that the observed distribution of forest, savanna and grassland vegetation in the tropical and boreal regions of the world represent alternative stable states for equivalent environmental conditions (Hirota et al, 2011;Staver et al, 2011;Favier et al, 2012;Murphy & Bowman, 2012;Scheffer et al, 2012;Xu et al, 2016). But recently Hanan et al (2014Hanan et al ( , 2015 suggested that the adopted classification and regression trees (CART) approach, used to produce the MODIS VCF tree cover estimates, introduced a systematic bias which makes the MODIS VCF product inappropriate for the analysis of % tree cover frequency distributions. This point was countered…”

“…Discontinuities in tree cover frequency distributions derived from this data have been used to support the hypothesis that the observed distribution of forest, savanna and grassland vegetation in the tropical and boreal regions of the world represent alternative stable states for equivalent environmental conditions (Hirota et al, 2011;Staver et al, 2011;Favier et al, 2012;Murphy & Bowman, 2012;Scheffer et al, 2012;Xu et al, 2016). But recently Hanan et al (2014Hanan et al ( , 2015 suggested that the adopted classification and regression trees (CART) approach, used to produce the MODIS VCF tree cover estimates, introduced a systematic bias which makes the MODIS VCF product inappropriate for the analysis of % tree cover frequency distributions. This point was countered by Staver and Hansen (2015) arguing (i) that the approach taken by Hanan et al (2014), using simulated EO data and pseudo satellite metrics to demonstrate that an artificial bias is generated by the CART approach, does not reflect the complexity and variability of landscapes and vegetation across the globe and (ii) that the CART model used by Hanan et al (2014) was highly pruned with very few nodes (nodes = 9) compared to that used for the MODIS VCF product (nodes = 109) resulting in a less smooth gradient of % tree cover values.…”

Inferring critical thresholds of ecosystem transitions from spatial data

“…The tree cover from MODIS VCF, however, is not well resolved below 30–20% cover (Hanan et al. , Abis and Brovkin ). While both NDVI and EVI are related to the amount of photosynthetic activity, EVI captures woody cover better because, unlike NDVI, it does not saturate at high values of photosynthetic activity (Glenn et al.…”

Resilience of tropical forest and savanna: bridging theory and observation

“…An obvious next step would now be to develop more detailed models that link the results of the long tradition of ground-based work (Scott et al, 2014) with the massive amounts of remote sensing data now available. While all remotely sensed data have important associated uncertainties (Roy et al, 2008;Hanan et al, 2014;Hanan et al, 2015), tree-cover and burned-area datasets show robust patterns over most of the tropics (Roy et al, 2008;Giglio et al, 2013;Hanan et al, 2015). Synthesizing this information with detailed observations of fire spread and tree mortality in relation to species traits and landscape geometries in spatially explicit and ecologically realistic models may bring us closer to a true understanding of the mechanisms that shape tropical landscapes in such surprisingly universal ways.…”

“…We inferred stable and unstable states of tree cover (minima and maxima in the potentials) for moving windows of the climatic variables (20 mm increments for MAP and 1% increments for MSI), where we applied Gaussian weights to the climatic variables (standard deviation was 5% of the entire range of the climatic variable) (Hirota et al, 2011). As the MODIS VCF product used here is not without bias Hanan et al, 2015), it precludes detailed interpretations for specific tree cover values. However, it has recently been validated that observed forest-savanna bimodalities are no artifact of the dataset (Staver & Hansen, 2015), implying that the bias is smaller than the signal.…”