Ecological Drought Adaptation

Climate change adaptation strategies are human responses to a climate change impact that have been designed to reduce harm, or take advantage of new opportunities, as the climate continues to change. Here we use “ecological drought adaptation strategies” to mean responses that specifically address drought impacts on ecological systems. Note that climate change adaptation strategies differ in intent from climate change mitigation strategies, which focus on reducing the drivers of climate change, and include efforts to reduce the use of fossil fuels that contribute carbon dioxide to the atmosphere. Ecological drought adaptation strategies will likely include actions and tools that are also incorporated into drought preparedness, drought response, and drought mitigation strategies. Tools and actions from drought planning are highly relevant - but we frame responses as climate change adaptation to highlight the role of climate change as an increasingly important factor shaping patterns of drought occurrence, drought intensity, and the potential for ecological transformations (e.g., a change from forest to grassland) following major climate-intensified drought events. Similarly, while not explicitly in the title, consideration of the role that human water use plays in reducing water availability for nature is an important aspect of “adapting” to drought risks under climate change. As described in our Ecological Drought Framework, strategies here emphasize reducing ecological risks, which we expect will benefit people by helping to maintain key ecological services. We expect this broadening of the scope of drought planning to address connections between ecological vulnerabilities and human well-being and water use, as well as climate change, will require updates to drought planning approaches. ecoDIVA is intended to help support groups in the process of making these updates, so communities can be better prepared for current droughts, and droughts of the future.

To be as effective as possible, an ecological drought strategy should directly address the most important drought driver(s). We expect that in many places, long term investment in managing ecological resources in appropriate ways, on appropriate sites, will reduce the risks of drought impacts to people and nature. However, we also expect increases in the severity of drought (higher probability of “megadroughts,” as well as more episodes of drought that exceed the adaptive capacity of the current system), where early warning of impacts may be the best we can hope for in terms of strategic response. While these ideas can be hard to accept, the need to think about thresholds is an important insight, as we are likely better off not investing limited resources in on-the-ground actions that are not likely to be successful. The key challenges are how do we evaluate the risks to the systems and services we care about, and evaluate the potential costs and benefits of action?

ecoDIVA presents maps and other visualization tools that describe watershed-specific relationships between tree health responses and a suite of drought drivers; variables with the largest “petal” are most strongly linked to the tree health response. Recall that these patterns were observed during a specific drought period, do not reflect all possible drought drivers, and often vary across watersheds - so they should be interpreted with these caveats in mind. For those tasked with conservation planning, management, or restoration of these forests, and for stakeholders interested in sustaining forest-based ecosystem services, a next step might be to identify adaptation strategies that can be deployed in response to the specific drought driver(s) that are most closely related to the tree health decline in a focal watershed. In the set of strategies below, drivers and responses are grouped according to our Ecological Drought Framework. We hope that connecting spatial data and strategies in this way helps groups of collaborators build a shared understanding of what is at risk, and how they can most effectively invest resources for management and restoration.

The ecological drought adaptation strategy list below is in development, and is primarily intended to illustrate the linkage between key drivers identified in the model and relevant actions; it is not exhaustive or prescriptive. Components of actions often overlap across the topical areas, reflecting the inherent connectedness of all of the factors influencing drought exposure, sensitivity, adaptive capacity, and impacts on people and nature. Similarly, a final management plan might incorporate concepts from multiple strategy types, reflecting the interaction between multiple drivers, and the idea that risks can often be reduced through the application of several management tools or siting decisions. Most notably, constraints related to identifying appropriate data for human water use in this example carries through to our presentation of responses - we lack adaptation examples for these forested watersheds that focus on ways to reduce impacts from human water use. In future case studies, we will expand our modeling breadth to include aspects of human water use that act as drivers of drought risk.

Our Upper Missouri Headwater case study model includes two variables we group as ecological characteristics, canopy cover and vegetation type. When one or both of these petals are long on the flower plot, this relationship suggests that management actions that influence forest characteristics at appropriate locations in the focal subwatershed can help reduce negative impacts on tree health under similar future drought conditions. We divide responses into several sets of actions below, but many actions under different subheadings overlap in potential actions and intent. In our Upper Missouri Headwaters case study model, our response variable is health of current trees, but we also include some approaches for forest restoration/management after major losses of forest trees in the list below.

  • Manage/restore the forest in ways that reduce water demand and/or incorporate approaches for increasing the adaptive capacity of current and future forests and ecological systems. If canopy cover is indicated as a drought sensitivity driver, consider selective harvest (tree thinning) or prescribed fire to reduce tree density, and reduce competition for water.
    • If thinning, consider the drought tolerance/evidence of drought stress in selection of trees to remove.
    • When restoring or managing systems, consider updating tools used to determine planting intensity, and/or consider deploying prescribed fire at early stages, to help promote a level of forest density that is sustainable as drought conditions arise.
    • If vegetation type is identified as a key driver, consider options for shifting species (or phenotype) composition through planting or selective harvest.
  • Anticipate impacts related to vegetation type: Consider how shifts in species distributions and changes in biodiversity currently occupying focal sites may influence forest communities and ecosystem services.
    • Anticipate vegetation change impacts on snowpack (as indicated by Snow Water Equivalent). Consider the potential for short term (loss of current coniferous forest cover), and longer term (conifer range shifts) and potentially major changes (species loss, system transformation) to alter the relationship between temperature, precipitation, tree cover and snowpack duration.
  • Prioritize protection of forested sites that have canopy cover characteristics and/or vegetation types that suggest to lower sensitivity to drought.
  • Additional options related to ecological characteristics that don’t currently link to a petal in our model because they focus on forest regrowth:
    • Evaluate planting methods used in restoration actions. If survival of young trees is likely to be a key bottleneck, consider how the choice of regeneration techniques (site preparation & seeding, vs. planting seedlings, or planting larger saplings) vary in terms of drought sensitivity & cost.
    • Consider options for saving tree seeds now to prepare for future situations where drought-influenced wildfires strongly reduce natural forest regeneration potential & reduce seed sources.

The case study model includes six variables that we group as landscape characteristics - all are derived at least in part from datasets on topography/elevation, which represent persistent attributes of site conditions that influence how solar radiation (insolation) and water availability vary across space. In addition to more generalized climatic conditions in a given time period, these location-specific factors influence a tree’s exposure and adaptive capacity relative to other trees in the same watershed. From simple to more complex, the landscape characteristic variables are aspect, slope, CHILI, TPI, TWI and CWD. To perhaps oversimplify, aspect and CHILI focus on a site’s exposure to solar radiation; slope, TPI, and TWI focus on how the shape of the land influences soil moisture; and CWD models landscape effects on water stress for plants, incorporating climatic and snowmelt data. See the modeling approach, and -buttons that accompany the variable maps for more information. In general, we found that the plant stress index CWD was the most important landscape variable “petal” in our focal watersheds. To interpret this index, it’s often helpful to look at maps of the other landscape variables. In the adaptation options below, we lump these variables together as indicators of the role of the landscape in driving variation in exposure and adaptive capacity, but try to highlight how to link these persistent features to relative risk of drought impacts. Note that if these petals were all small for a watershed, that suggests that topography was not a strong contributing factor in the focal drought - but these actions may promote forest health and diversity under more average conditions.

  • Manage/restore the forest in ways that recognize the contributions of topography and elevation as drivers of exposure, and the adaptive capacity of forests. This idea builds on centuries of life history and ecology research - for example, we expect species with higher drought tolerances to do well on south facing slopes.
    • As we look to the future, our understanding of how species are distributed across slope, aspect, and soil moisture gradients can inform management and restoration - if a site is too dry for specific tree species now, are there other sites on the landscape with soils and topography that may be more suitable?
    • Where tree density values are similar in the case study maps, evaluate the model results to see if landscape variables indicate where this level of density appears to be too high. If the areas where tree health remained high align with cooler aspects and higher soil moisture values, this can help inform discussions of where to thin or apply fire.
  • Incorporate consideration of topography into protection decisions.
    • Areas with more variation in topography are likely to have more variation in microclimates, and can potentially support a higher diversity of species on a per-area basis.
    • Prioritize protection of forested sites in lower drought risk locations (drought refugia) - sites with cooler aspects, soils with higher moisture-holding capacity, locations where snow is likely to persist longer into the spring/summer.
    • Consider the role that drought refugia can plan in reducing risks related to payment for ecosystem-service based strategies that help support forest management and protection. If you are designing a forest protection project using carbon credits or a waterfund, anticipate how drought risk varies across the landscape, and choose lower risk sites to ensure these services will persist.
  • Anticipate impacts to freshwater ecosystem services: If water from snowpack is a key value, protect those places where the shape of the landscape and tree cover interact to sustain snowpack.

These strategies connect to the petal for the different drought indices, which represent or incorporate climatic variables. In general, if the dominant drought driver (indicated by a longer petal) for a given watershed is a climate factor like evaporative drought demand (EDDI), this suggests that climatic exposure factors overrode factors like forest characteristics, site conditions, and the adaptive capacity provided by topographic factors that modify soil moisture, and vegetation-climate interactions. Besides the long-term strategy of working on climate change mitigation (including planting trees), in many cases we don’t have on-the-ground actions for effectively reducing risks when climatic factors dominate other drivers. These are the cases noted in the introduction where what we are learning is that on-the-ground actions like thinning would be less likely to help keep trees from crossing critical water needs thresholds if another drought of this magnitude were to occur in these watersheds. Instead, many of the responses here focus on monitoring and anticipating impacts, and planning drought response actions (i.e., actions that can be taken during or after drought impacts occur).

  • Manage/restore the forest in ways that reflect key drought indicators. If Snow Water Equivalent (SWE) and/or soil moisture (LERI) were important drivers of the ecological drought response, consider ways to keep snowpack and moisture on the landscape longer in appropriate sites.
    • Invest in planting or protecting conifers that insulate snow.
    • Consider structures that mimic the behavior of beavers and increase flooding/soil moisture so that soil moisture stays higher for more of the growing season.
    • What are the factors that best predict variation in impact when climate drivers are the most important factor? Are there any places that seem to act as refugia? Are there topographic or other persistent factors (e.g.,, soil texture, or presence of a water holding layer in soil) that indicate these locations?
  • Anticipate how trends in climate drivers that relate to drought risk are likely to lead to impacts, including shifts in species distributions and changes in biodiversity.
    • Identify the suite of climate datasets and indices that are likely to provide insight into future conditions and risks. Which drought indices (EDDI, FDSI, SPI, LERI) work best for the ecological impact you care about?
    • Develop ways to communicate the changes in a subset of these indicators to the public to help broaden support for pro-active management.
    • Consider the potential for minor changes (range shifts) and potentially major changes (species losses, system transformation) in forest species may affect biodiversity and ecological services. Is there information available to help predict how those shifts may occur (upslope or downslope; key mechanisms for species loss like insect/fire interactions)? These kinds of discussions can help identify which indices and ecological variables to monitor.
  • Connect the drought indices noted above to plans for responding to drought. Develop “theories of change” that link drought indices and other climate data to tangible ecological, social, and hydrologic impacts and responses. These can be hypotheses that require testing, but invest time in connecting indicators to specific impacts so that monitoring and research can be developed to inform future decisions. With this foundation in place, when monitoring indicates that ecological drought impacts are likely, plans can guide actions to:
    • Reduce water withdrawals for irrigation and other human uses.
    • Close access to fire-prone areas to reduce the likelihood of ignition.
    • Plan post-drought and post-fire planting efforts that are designed to reduce future drought and fire vulnerabilities (e.g., by incorporating more drought- and fire-tolerant species).
    • Plan post-fire erosion control actions.

In this case study, we did not have access to data on human water use that we could incorporate into the tree health model. Potentially, water uses such as irrigation for agriculture, snowmaking for ski resorts, or use of surface or groundwater for human or industrial water needs could reduce water available for ecological systems. Potentially, irrigation or snow-making could also have local beneficial impacts if additional water is brought to the systems through human actions (likely with negative impacts in other locations, or at a future time as aquifers become depleted). A long petal for human water use would suggest that these uses were playing an important role in determining whether ecological systems past a key threshold. Actions that could link to these petals would depend on the specific form of water use, but could include limiting water withdrawals, changing the timing of these withdrawals, and related strategies for reducing the chances that human water uses would exacerbate the impacts of climatic, ecological, and landscape factors in driving drought risk.