Sixteen million lidar shots against the idea of one pantropical carbon curve

Researchers affiliated with the Smithsonian Tropical Research Institute, the University of Maryland, NASA and Brazil's Universidade Federal dos Vales do Jequitinhonha e Mucuri analysed 16 million NASA-GEDI satellite estimates of aboveground biomass from 2020 across the Amazon, the Congo Basin and Southeast Asia, and reported in Nature that temperature, aridity, soil nutrients and other environmental variables act on biomass differently from place to place ^[1][2]. If your carbon accounting rests on one fitted relationship between climate and biomass applied across the tropics, that is a specification error rather than a footnote ^[1].

Take the two headline gradients. Hotter sites generally held less biomass, but the sensitivity ranking was Congo Basin (particularly sensitive), then the Amazon (moderately), then Southeast Asia (relatively insensitive) ^[3]. Water limitation inverted the order: it mattered most in Southeast Asia, where biomass fell sharply at more arid sites, while Amazonian forests peaked at intermediate aridity and African forests were relatively insensitive to aridity ^[4]. Those three aridity responses are not three settings of one curve. A monotonic decline, a hump, and a flat line cannot be reproduced by a single pantropical function, so any pooled fit is mis-specified in at least two of the three basins ^[1]. On temperature, a pooled coefficient is a weighted average of the regional ones, which means it necessarily overstates Southeast Asia's response and understates the Congo Basin's ^[2].

The modifiers make pooling worse, not better. Climate effects were themselves conditioned by soils and landscape features across all regions ^[5], and in the tallest forests, where trees exceed 70 metres, the authors suggest storm damage from lightning and windthrow may be the dominant drag on biomass ^[6]. Structure is not incidental either: lightweight-wood trees such as balsa store far less carbon than dense hardwoods like rosewood or ebony, and which of those a hectare contains depends on soil and whether the ground is flat, mountainous, dry or swampy ^[13].

What this settles is worth separating from what it does not. The authors argue the results resolve earlier conflicting findings on biomass patterns by showing the disagreement partly reflects real biological variation rather than methodological differences ^[7]. Co-author Helene Muller-Landau of STRI says the study "definitively shows that tropical forests on different continents respond differently to climate" ^[8], and attributes the divergence to enduring legacies of Africa's drier and Southeast Asia's wetter climatic histories ^[9]. STRI paleobiologist Carlos Jaramillo makes the same point on geological timescales ^[12]. Lead author Matheus Nunes of Maryland goes further on method, arguing global analyses are not enough and that local analysis and local experts are needed to predict climate effects ^[10].

Two limits operators should hold onto. The published summary reports orderings of sensitivity, not slopes, so it tells you a pooled model is biased without telling you by how much ^[3][4]. And this is a single year of footprints compared across sites, a spatial substitution for time rather than an observed response to warming ^[3].

Watch whether biomass map producers and methodology writers move to region-specific fits, and watch the plot networks that would validate them: the source flags GEO-TREES as a new network measuring forest carbon with consistent methods across sites ^[14]. GEDI itself is the constraint that made the comparison possible, using laser pulses to tie three-dimensional structure to climate, soils and topography ^[11].