Variable species establishment in response to microhabitat indicates different likelihoods of climate-driven range shifts

AbstractClimate change is causing geographic range shifts globally, and understanding the factors that influence species’ range expansions is crucial for predicting future biodiversity changes. A common, yet untested, assumption in forecasting approaches is that species will shift beyond current range edges into new habitats as they become macroclimatically suitable, even though microhabitat variability could have overriding effects on local population dynamics. We aim to better understand the role of microhabitat in range shifts in plants through its impacts on establishment by Q1) examining microhabitat variability along large macroclimatic (i.e., elevational) gradients, Q2) testing which of these microhabitat variables explain plant recruitment and seedling survival, and Q3) predicting microhabitat suitability beyond species range limits. We transplanted seeds of 25 common tree, shrub, forb, and graminoid species across and beyond their current elevational ranges in the Washington Cascade Range, USA, along a large elevational gradient spanning a broad range of macroclimates. Over five years, we recorded recruitment, survival, and microhabitat (i.e., high resolution soil, air, and light) characteristics rarely measured in biogeographic studies. We asked whether microhabitat variables correlate with elevation, which variables drive species establishment, and whether microhabitat variables important for establishment are already suitable beyond leading range limits. We found that only 30% of microhabitat parameters covaried with elevation. We further observed extremely low recruitment and moderate seedling survival, and these were generally only weakly explained by microhabitat. Moreover, species and life stages responded in contrasting ways to soil biota, soil moisture, temperature, and snow duration. Microhabitat suitability predictions suggest that distribution shifts are likely to be species-specific, as different species have different suitability and availability of microhabitat beyond their present ranges, thus calling into question low-resolution macroclimatic projections that will miss such complexities. We encourage further research on species responses to microhabitat and including microhabitat in range shift forecasts.
MethodsStudy site We utilized the large macroclimatic and elevational gradients of the Cascade Range in Washington, USA to transplant seeds along each of 2 large transects spanning ~1200 m on the West and East side of the Cascade Crest on the traditional lands of the Nlaka’pamux, Nooksack, Okanagan, and Methow peoples. Both transects are characterized by topographically complex terrain and differ substantially in their macroclimatic characteristics. The West transect (Mount Baker National Forest) is warmer and wetter than the East transect (Okanagan National Forest). Both transects are characterized by montane to subalpine species common in the Pacific Northwest, with an abundance of cedars, firs, heathers, and understory forbs and graminoids. We selected 15 sites along each of the 2 transects (n = 30 sites) using satellite images of accessible areas, and identifying areas (i.e., blocks) that had both low and high tree canopy openness. In the field, we established two blocks per site (n = 60 blocks) to encompass different levels of canopy openness, with one block in the relatively most open area and the other block in the relatively most closed canopy. Within each block, we selected three replicate plots separated by ~2 m (n = 180 total). Each plot contained three 50 cm x 50 cm quadrats ~ 20 cm from one another (n = 540 total). To ensure substrates had recruitment potential, we selected quadrats that had no saplings >10 cm in diameter, no large rocks that covered >10% of the quadrat, and ≥ 5% cover from seedlings or understory species. Of the three quadrats, we added seeds to two (separating confamilal pairs to separate plots to facilitate field identification of seedlings) and left the third quadrat unmanipulated to control for background recruitment. We grouped species by climate affinity (see next section) and sowed one quadrat per plot with seeds of species that generally occur in cooler temperatures and the other quadrat with seeds of species that generally occur in warmer temperatures. Species data We sowed 25 native species encompassing a variety of functional groups, climate affiliations, seed sizes, and regional prevalence to capture a variety of species responses in the community. Species included trees (Abies grandis, A. lasiocarpa, Picea sitchensis, P. engelmannii, Pinus ponderosa, P. contorta), shrubs (Mahonia nervosa, M. aquifolium, Rubus ursinus, R. spectabilis, Sambucus cerulea, S. racemosa, Sorbus sitchensis, Vaccinium parvifolium, V. deliciosum), forbs (Eriophyllum lanatum, Anemone occidentalis, Erigeron perigrinus, Lupinus latifolius, Maianthemum dilatum, M. racemosum, Tolmiea menziesii, Tellima grandiflora), and graminoids (Carex stipata, Carex...