Climatic microrefugia under anthropogenic climate change: implications for species redistribution

Figure showing the spatial distribution of the maximum of canopy height (m) derived from airborne LiDAR data at 50 cm resolution across the forest of Compiègne in northern France. Copyright by Tarek Hattab (EDYSAN: https://www.u-picardie.fr/edysan/).

#E4 award #Runner-up

by Jonathan Lenoir, Tarek Hattab and Guillaume Pierre

If exterior temperature increases steadily to the point that you cannot stand it anymore, one option is to pack your bag and move to a cooler place that suits you until temperature goes beyond your tolerance threshold again meaning that you (or your descendants if you are too old already to endure the journey) have to move again to another cooler place. Alternatively, you (and your descendants) can stay at a shelter where interior temperature remains steady over time – like a cave where you would keep your best wine – and wait until exterior temperature are suitable again to explore the neighbourhood. Let us call this shelter a “microrefuge” (plural: microrefuges) if temperatures are suitable again within your lifetime and a “microrefugium” (plural: microrefugia) if temperatures are not going to be suitable again for what seems to be a very long long time to you and your descendants (i.e. several generations).

Now replace “you” by any kind of living organism. To avoid extinction under anthropogenic climate change, a species may either shift its geographical range over time or persist within microrefugia, assuming that conditions are changing too fast to allow acclimation or microevolutionary processes to happen. Given this, identifying microrefugia and assessing their capacity for long-term persistence of biodiversity threaten by anthropogenic climate change has recently been highlighted as a daunting but timely endeavour, especially so in lowland agricultural landscapes where both habitat fragmentation and distance between isotherms hinder species range shifts. Current species distribution models (SDMs) and their projections under climate change are based on coarse-resolution climatic grids that do not account for interior microclimatic processes such as climatic stability (buffering and decoupling) over time and climatic heterogeneity occurring over few metres. This spatial mismatch between the SDMs’ spatial resolution and the spatial resolution at which microclimatic processes happen makes the identification of climatic microrefugia challenging.

The most recent scientific advances in modelling microclimate at fine spatial resolution have already proposed to incorporate physiographic processes (e.g. cold-air pooling) generated by topographic features over few metres but these do not yet account for biophysical processes (e.g. buffering and decoupling) generated by canopy cover and the vertical structure of vegetation. For instance, the forest understory, depending on the level of canopy density, is well known for being relatively buffered from exterior climatic conditions (cf. cooler/warmer conditions during summer/winter time inside than outside a forest stand), especially so within ravine forests . Although this knowledge is pretty strong in the scientific literature, none have incorporated it in a process-based modelling approach to account for the buffering and decoupling capacities of forest ecosystems. Yet, we can now benefit from very-high-resolution (VHR) digital elevation model (DEM) as well as VHR digital canopy model (DCM) thanks to small-footprint light detection-and-ranging (LiDAR) technology, being airborne (figure above) or terrestrial LiDAR. In the meantime, temperature sensors have been miniaturized and popularized to the point that these tiny loggers are now widely used in field ecology to record in-situ temperature conditions (figure below). Thus the combined use of VHR DEM, VHR DCM and temperature loggers has never been so timely to model sub-canopy temperatures at fine spatial resolution (cf. figure below) and thus improve projections of the spatial redistribution of forest-dwelling species.

Picture of a iButton mounted at 5 cm above the ground to record temperature conditions every 2 hours. Copyright by Jonathan Lenoir (EDYSAN: https://www.u-picardie.fr/edysan/).

Here, we propose a spatially hierarchical downscaling framework combining a free-air temperature grid at 1 km resolution, a DEM at 25 m resolution and LiDAR data at 50 cm resolution with knowledge from the scientific literature to mechanistically model sub-canopy temperatures and account for microclimatic buffering and decoupling in SDMs. Finally, we provide a working agenda with future guidelines to better incorporate microclimate and microrefugia into SDMs. For instance, we recommend the use of well-designed networks of temperature loggers across large spatial extents to refine the parameters (retrieved from the scientific literature here) that we used here to mimic the buffering and decoupling processes of climatic microrefugia. The way we accounted for these key processes in our modelling framework is very simplistic but realistic and we hope that it can be used as a start to better account for microrefugia when projecting species redistribution under anthropogenic climate change.

Figure displaying free-air temperature at 25-m resolution and sub-canopy temperature at 50-cm resolution within the forest of Compiègne in northern France. Copyright by Tarek Hattab (EDYSAN: https://www.u-picardie.fr/edysan/).

By favouring in situ species persistence, climatic microrefugia will not only provide a safe haven for biodiversity but also ensure the long-term provision of regulating ecosystem services, thus changing our view of the consequences of climate change on living biota. Recognizing the ecological relevance of climatic microrefugia for species redistribution under anthropogenic climate change will open new avenues for adaptation management in conservation biology.

Comments