Scaling up: How forest hydraulic diversity may throw off global climate models

Featured Image: An eddy flux tower in Waccamaw Neck, South Carolina, Source: Scott Allen

Reference: Anderegg, William R.L., A.G. Konings, A.T. Trugman, K.Yu, D.R. Bowling, R. Gabbitas, D.S. Karp, S. Pacala, J.S. Sperry, B.N. Sulman, and N. Zenes. 2018. Hydraulic diversity of forests regulates ecosystem resilience during drought. Nature 561, 538-541.

Global Land Surface Models

Global land surface models provide a mathematical description of how the physical, chemical, and biological processes of the land surface interact with those of the atmosphere. As you can imagine, the make-up of the land surface, whether it be ocean, forest, urban, or grassland, has a large effect on the exchange of water, carbon, and energy between the land surface and the atmosphere. Thus, global land surface models play an important role in global climate models.

Because actual land surface processes are complex and not entirely understood, global scale models rely on the simplification of reality in order to be feasible. In other words, scientists cannot account for the variability in water and energy fluxes (the rate of flow between the land and the atmosphere) for two adjacent plants. They can only estimate the fluxes of those plants in aggregate by combining kilometers of land and estimating total flux from that entire land area. Unfortunately, a large amount of information is lost when performing this type of analysis, and some of that variability may explain why climate predictions are so uncertain. Let’s dive into why some of these uncertainties exist by picking apart some of the pieces that make up a land surface model.

Vegetation representation

Vegetation has traditionally been classified using biome divisions (tundra, arid desert, subtropical rainforest, etc.) or more recently, plant functional types. Plant functional types are usually generated by identifying the dominant plant type for the land area and assigning values for key characteristics such as the potential for reproduction, competitiveness, and how carbon is allocated. In many cases this method not only fails to include the variability inherent in the land area being represented but it often does not include important vegetation characteristics such as the various ways the different plant species control water movement to avoid water limitation (hydraulic strategies).

Plant species have a huge range of hydraulic strategies to prevent drying during periods of water limitation. Some plants close their stomata (the small holes in leaves where carbon dioxide enters and water exits the plant) at the first hint of water limitation, whereas others can maintain open stomata to much lower soil moisture availability or have access to deeper groundwater via long roots. Unfortunately these various strategies are not well represented by different plant functional types used in global land surface models. This can lead to large deviations between observed fluxes and modeled outcomes.

Land surface and atmospheric coupling

So how do plants actually interact with the atmosphere? To really understand this concept you need to have a better idea of the surface energy budget. The amount of energy reflected and absorbed by the earth’s surface depends on the surface material. For instance, lighter surfaces like ice reflect more light, whereas darker surfaces such as a forest canopy can absorb more light and thus heat up faster. Some of the energy that hits the earth’s surface is used to convert liquid water to water vapor through two processes, evaporation and transpiration. Transpiration is the process by which vegetation uses solar energy to convert carbon dioxide into biomass via photosynthesis and releases water vapor to the atmosphere in return. Evaporation occurs at the land surface when soil moisture vaporizes and is released into the atmosphere. The two processes are often combined together as evapotranspiration and known as latent energy flux in the surface energy budget. Latent energy is energy used in phase changes rather than temperature changes (sensible heat flux).

Transpiration is primarily controlled by the vapor pressure deficit (VPD) or the difference between the instantaneous amount of moisture in the air and the amount of moisture in the air when it is fully saturated with water vapor (humidity gradient). When the VPD is high, the atmosphere pulls water upward through the trees from the soils. However, as the pull from the atmosphere increases (either due to less water in the soils or in the atmosphere), plants begin to tighten their stomata and regulate flows. When the soil moisture drops below a certain point, the tree can close its stomata or the opening in leaves where gas and water exchange occurs to prevent further water loss. This is where biology can make big differences in climate models.

Hydraulic diversity

A recent study investigated how large of a role site variability may play in controlling a forest’s drought resilience. The idea is that forests that contain species with many different hydraulic characteristics will be more resilient in the face of drought as compared to sites that have lower hydraulic diversity. A higher diversity in hydraulic strategies may enable forests to be more efficient with water use at the landscape scale. For instance, one species may start to access deeper soil water while another species may reduce transpiration rates. This leaves more water for other species that may have less hydraulic control.

Fig. 1: Flux tower, Waccamaw Neck, South Carolina. Source: Scott Allen

William Anderegg from the University of Utah’s School of Biological Sciences and colleagues used data from 40 flux towers across global temperate and boreal forests. Flux towers extend above plant canopies to measure the total amount of carbon and water that is traveling down from the atmosphere and up from the land surface (Cover photo and Fig. 1). The sum of the these carbon and water values allows scientists to calculate the net exchange between the surface and the atmosphere. Data from these towers are used to calibrate and validate global land surface models as well as better understand how plants regulate gas and energy exchange between the earth’s surface and atmosphere.

At each tower, the authors looked at the relationship between latent energy exchange, a proxy for transpiration, VPD, and soil moisture during periods of water limitation. A better relationship between latent energy exchange, VPD, and soil moisture equated to higher sensitivity to drought. In other words, as VPD increased and soil moisture decreased, plants responded more strongly by reducing transpiration. On the other hand, if plants had a higher ability to control water use, the overall site transpiration would not be as strongly related to VPD or soil moisture. These sites were more resilient in the face of drought. The strength of this relationship determined drought sensitivity in the study (Fig. 2).

Fig. 2: Relationship between latent energy exchange and VPD shows how drought sensitivity increases with the slope of the relationship as defined by the study. Source: Mary Grace Lemon

The team investigated many different site variables to determine their influence on drought sensitivity such as forest age and specific leaf area index or the ratio of leaf area to dry biomass. However, the site characteristic that explained the largest amount of variability in the data was the standard deviation of the hydraulic safety margin (HSM). HSM is defined as the difference between naturally occurring xylem (tree tissue that conducts water) pressures, and pressures that would cause hydraulic failure or inability of the xylem to transmit water from the roots to the leaves. In other words, sites with higher hydraulic diversity were more resilient against drought. Not accounting for this hydraulic diversity in global land surface models could potentially cause over or underestimation of the amount of water and carbon being transported between the land and atmosphere.  

There is big push to better represent tree physiological processes in global land surface models. Variability in water use strategies is a good place to start, but the exact mechanism controlling this observed pattern is still unknown. This makes it difficult to come up with a relatively simple way to represent this behavior on a global scale. This paper highlights the importance of looking past average values and respecting the enormous amount of variability that exists in plant physiology even within a single forest stand.

Share this:
Mary Grace Lemon

Mary Grace Lemon

I am an ecohydrologist by training with special interests in wetland systems but my interests and experiences are diverse. I have worked in a variety of ecological systems including coastal estuaries, riverine wetlands, and forested ecosystems. Outside of research and management work, I like to spend most of my time outside doing various recreational activities generally with a water component.

Leave a Reply