How deep should we go to understand roots at the top of the world?

Informed by vegetation maps across high-latitude landscapes, terrestrial biosphere models are a tool that can be used to predict changes in the composition and function of vegetation, above- and belowground, across the land surface in response to changing environmental conditions. However, terrestrial biosphere models represent vegetation characteristics at a finer grain than mapped vegetation communities. These models group plant species that colonize high-latitude biomes by their functional trait variation into plant functional types (PFTs) that characterize the impacts of plant species on, and their response to changes in, their surrounding abiotic and biotic environment. Blume-Werry et al. (2023) found that vegetation mapping units that broadly incorporate multiple plant species and functional types are too coarse, or encompass too much biological variation, to fully capture belowground plant trait variation. However, they did find that they could successfully cluster rooting depth observations into ‘Root Profile Types’, suggesting that modeling PFTs may be a useful tool to characterize above- and belowground linkages across high-latitude environments.

In many arctic and boreal ecosystems, plant roots are constrained by permafrost to a shallow ‘active layer’ of soil that thaws progressively over the course of each growing season. Blume-Werry et al. (2023) identified active layer thickness and the closely related minimum temperature of the coldest month as two of three main abiotic drivers constraining rooting depth distribution in their analysis (a third, cation exchange capacity, is more indicative of nutrient availability than a physical impediment). Furthermore, waterlogging can limit root distribution to surface, oxic soils, and can lead to a thick layer of poorly decomposed, organic peat at the soil surface with different characteristics from mineral soils (Fig. 1; Walker et al., 2003). Indeed, Blume-Werry et al. (2023) found that despite similarities in species composition between wetland and graminoid tundra in CAVM mapping units, rooting depth in wetland tundra was shallower than graminoid tundra. This may indicate that waterlogged conditions can constrain rooting depth distribution, even in vegetation communities dominated by species with aerenchymatous roots. Ranging from rootless mosses and plant-like lichens to vascular graminoids and shrubs, and deciduous and evergreen trees, PFTs inhabiting the arctic tundra and boreal forest vary in their rooting depth distributions, their interactions with soil microbiota, and their ratio of belowground to aboveground tissues (e.g. root : shoot ratio; Chapin et al., 1996). However, terrestrial biosphere models have often neglected the unique characteristics of the species that colonize high-latitude biomes, especially belowground (Iversen et al., 2015, 2018). While Blume-Werry et al. (2023) did not find that CAVM vegetation classifications could be used to predict and scale rooting depth distribution across the pan-Arctic, their approach can be further refined to link tundra plant rooting depth distribution with surrounding soil characteristics in models (Drewniak, 2019).

The capability to predict belowground form and function across the vast expanse of the northern high latitudes is a tangled problem that spans spatial and temporal scales and encompasses biological and environmental considerations. The authors propose moving forward by considering ‘root functional types’. Here, we suggest a few additional paths forward so that we as a belowground research community might find the connections that unlock our understanding and prediction of belowground processes in a rapidly changing region of the world.

Blume-Werry et al. (2023) took on the important and difficult challenge of predicting root form and function from observed aboveground vegetation community composition across the pan-Arctic. The inability to capture large-scale variation in the depths of roots, and ensuing differences in modeled carbon emissions, with vegetation maps by Blume-Werry et al. stems from multiple possible sources, not least of which is that these maps are of aboveground vegetation. A solution proposed by the authors is to categorize ecosystems by their belowground, rather than aboveground, features. Understanding the role of the belowground in arctic and boreal ecosystems may require a belowground-focused approach coupled with remote sensing (Blume-Werry et al., 2023; Yang et al., 2023), mechanistic modeling, machine learning (Langford et al., 2019; Shi et al., 2021; Sulman et al., 2021), and cross-disciplinary collaboration with empirical researchers (Sulman et al., 2021; Blume-Werry et al., 2023).