Moisture, thermal inertia, and the spatial distributions of near-surface soil and air temperatures: Understanding factors that promote microrefugia

Highlights

• Moisture reduces exposure to temperature extremes, facilitating climate refugia.

• Variability of air temperatures related to humidity, but soil temperatures to VPD.

• Moisture differentially affects spatial distribution of soil and air temperatures.

• Microrefugia will be more apparent under dry conditions.

Abstract

Climate change poses significant threats to biodiversity, but some species may be able to escape its effects in small locations with unusual and stable climates (microrefugia). However, there are still great uncertainties about where microrefugia are located, and the exact role that moisture plays in buffering extreme temperatures. In this study we quantified the effects of moisture on the distribution and variability of near-surface soil and air temperatures. We collected hourly 1 cm soil and 5 cm air temperatures and humidities at 111 sites from May 2011 to March 2012. Sites were diverse in terms of elevation (2–1428 m), distance from coast (180 m–403 km), canopy cover (0–100%), topographic exposure, and susceptibility to cold air drainage. We found that variability (diurnal range) of both soil and air temperatures decreased under moister conditions. While air temperatures were related more strongly to humidity, soil temperatures were related more to vapour pressure deficit (VPD). That is, both high temperature and low humidity were required before the VPD was sufficient to dry out the soil and allow soil temperatures to vary. We then used a regional regression approach to model the spatial distribution of minimum and maximum air and soil temperatures for each day over the 10 months in terms of latitude, elevation, canopy cover, distance to coast, cold air drainage potential, and topographic exposure to the south and northwest. We found that elevation was the dominant factor explaining the distribution of soil and air temperatures under moist conditions. Other factors, such as canopy cover and topographic exposure, had a stronger influence on air temperatures whenever humidity was low. However, these factors only affected soil temperatures at times when higher temperatures combined with low humidity to produce higher VPD. Our results provide new insights into how moisture influences the spatial distribution of near-surface soil and air temperatures. Microrefugia will be more apparent under drier conditions, but climate change may affect refugia for soil and air temperatures differently. Higher temperatures will cause VPD to increase more than would be expected by any change in humidity, and refugia in terms of soil temperatures may therefore become increasingly apparent.

Introduction

It has been predicted that climate change will cause latitudinal and elevational shifts in species distributions and lead to the possible extinction of 15–37% of species (Hughes, 2000, Thomas et al., 2004). However, these predictions may be overly pessimistic, in part because they are based on macroclimatic conditions and ignore small locations with unusual climates (microrefugia) where species may be able to escape the effects of climate change (Rull, 2009, Ashcroft, 2010, Dobrowski, 2011). Indeed, while ice sheets or extreme aridity may render large regions uninhabitable and force local extinctions or broad-scale migrations, phylogeographic evidence suggests that in unglaciated regions, extinctions and range shifts have been rare and most species have been able to persist through Pleistocene climatic cycles in microrefugia within heterogeneous landscapes (Byrne, 2008, Rull, 2012). Topographic heterogeneity is also recognised as an important buffer against future climate change (Ackerly et al., 2010, Scherrer and Körner, 2011), but we are only just beginning to identify the locations of microrefugia within topographically complex regions and understand how they function (Dobrowski, 2011, Ashcroft et al., 2012, Keppel and Wardell-Johnson, 2012).

There is still a lot of confusion around the exact definition and location of microrefugia (Ashcroft, 2010, Dobrowski, 2011, Keppel and Wardell-Johnson, 2012), and more effort is needed to determine the factors and processes that determine their location (Hampe et al., 2013). An important step towards this goal is determining what causes locations to experience climatic buffering (reduced variability and susceptibility to extremes). Buffering of near-surface temperatures may result from thermal inertia of moist soils (Lu et al., 2009), reduced net radiation fluxes due to canopy cover (Geiger, 1971), or topography that promotes cold air pools or creates shelter from winds (Ashcroft et al., 2009, Dobrowski, 2011; but see also Ashcroft et al. (2012) for evidence that cold air pools actually have higher climatic variability). This buffering of temperatures may cause the near-surface air at these locations to be decoupled from the free atmosphere under some weather conditions (Daly et al., 2010, Pepin et al., 2011), but this decoupling is not present under all weather conditions, and factors such as cloud cover, radiation, rainfall and wind can act across all sites simultaneously and cause some synchronicity in conditions between both coupled and decoupled sites. We argue that climatic buffering needs to be understood not just in terms of decoupling from the atmosphere, but also in terms of the spatially and temporally variable effects of a number of other climate-forcing factors (wind, radiation, soil moisture, cold air drainage potential, canopy cover etc.).

The processes that influence climatic buffering (e.g. radiation, cold air drainage, thermal inertia of soils) generally operate within a few centimetres of the soil surface, and hence localised buffered climates are more likely to be apparent at the soil surface (Chen et al., 1999). However, most meteorological observations are made at a standardised height of 1.5–2 m, and wind and convection will obscure localised climates at this height. It is difficult to convert standardised observations to near-surface temperatures, as relationships between the two are affected by factors such as cloud cover, time of day, season, wind, canopy cover, topographic exposure and proximity to coast (Wolfe, 1945, Geiger, 1971, Bond-Lamberty et al., 2005, Likso, 2006, Ashcroft and Gollan, 2012). To improve our understanding of climatic buffering and the potential impacts of climate change, it is important that studies focus on near-surface air and soil temperatures rather than standardised observations (Graae et al., 2012; but see also Ashcroft et al., 2009, Holden et al., 2011 for evidence that topoclimatic data may be linked with standardised observations, as this may allow us to produce topoclimatic maps of near surface temperatures in other places or times).

The zone a few centimetres above and below the soil surface is also of primary ecological importance for germinating seeds, sensitive saplings, foraging animals, bushfire fuel moisture and ecological processes such as decomposition, soil respiration and evaporation (Kennedy, 1997, Chen et al., 1999, McVicar et al., 2007, Kustas and Anderson, 2009, Holden and Jolly, 2011, Graae et al., 2012). For example, plants may successfully reproduce only at favourable places and times (their regeneration niche; Ranieri et al., 2012), and the spatial distribution of saplings may be restricted to localised environments rather than the entire species’ distribution (McLaughlin and Zavaleta, 2012). If these localised environments provide safe havens that are crucial for the persistence of populations, then accurate, fine-scale climatic data will be needed to understand and predict current and future distributions. As long as the quality of climate data continues to be overlooked as a source of error in species distribution models (Soria-Auza et al., 2010), and insufficient research is undertaken to relate global climate change to ecologically relevant microclimates (Kennedy, 1997), we will be unable to predict future impacts to microclimate-sensitive species (e.g. Roslin et al., 2009).

The magnitude of the differences between surface conditions and standardised observations should not be underestimated. For example, on a hot summer day, standardised observations of 40 °C may correspond with ground temperatures of 60 °C (Campbell and Norman, 1998, Ashcroft and Gollan, 2012). This introduces a bias that poses a fundamental problem when we compare, for example, thermal tolerances of species against standardised observations. It has been noted that warming tolerances (differences between a species critical thermal maximum and the mean temperature of warmest quarter at standardised observation height) can be up to 45 °C (e.g. Diamond et al., 2012), but a large portion of this is likely due to the differences between standardised observations and near-surface conditions rather than an actual buffer against warming (a portion is also due to the difference between mean temperatures and maximum temperatures). The fact that warming tolerances differ between species in arid and forested environments (e.g. Diamond et al., 2012) is also likely to reflect the different biases between standardised observations and near-surface conditions in these environments, which will obscure any actual differences in their exposure to warming.

Near-surface climatic conditions are also of interest because they are a crucial component of regional and global climate models (RCMs and GCMs), which are the basis of future climate predictions. For example, soil moisture affects the proportion of radiation that results in evaporation instead of temperature rise, results in feedbacks that prolong regional droughts or floods, and forms a crucial component of the planetary boundary layer that drives the underlying climate models (Evans et al., 2011). More research is needed to determine how changes in near-surface soil moisture and climatic conditions will modulate regional climates, and this may also improve our ability to downscale future climate predictions to finer scales (Diffenbaugh et al., 2005, Evans et al., 2011).

The objective of this study was to improve our understanding of how the variability and spatial distribution of near-surface soil and air temperatures are influenced by moisture. The specific goals were: (1) to quantify the variability of soil and air temperatures over short time-scales (diurnal range) and determine the relationships with humidity and vapour pressure deficit (VPD); and (2) to quantify how the spatial distributions of soil and air temperatures change over time, and explain these temporal trends in terms of how the effects of different climate forcing factors change under different moisture levels. To maximise the generality of results, we made observations over a large region with a diversity of habitats and topographic positions and a broad range of both elevation and distance to coast.