Guidelines and considerations for designing field experiments simulating precipitation extremes in forest ecosystems

Abstract

Precipitation regimes are changing in response to climate change, yet understanding of how forest ecosystems respond to extreme droughts and pluvials remains incomplete. As future precipitation extremes will likely fall outside the range of historical variability, precipitation manipulation experiments (PMEs) are critical to advancing knowledge about potential ecosystem responses. However, few PMEs have been conducted in forests compared to short‐statured ecosystems, and forest PMEs have unique design requirements and constraints. Moreover, past forest PMEs have lacked coordination, limiting cross‐site comparisons. Here, we review and synthesize approaches, challenges, and opportunities for conducting PMEs in forests, with the goal of guiding design decisions, while maximizing the potential for coordination. We reviewed 63 forest PMEs at 70 sites world‐wide. Workshops, meetings, and communications with experimentalists were used to generate and build consensus around approaches for addressing the key challenges and enhancing coordination. Past forest PMEs employed a variety of study designs related to treatment level, replication, plot and infrastructure characteristics, and measurement approaches. Important considerations for establishing new forest PMEs include: selecting appropriate treatment levels to reach ecological thresholds; balancing cost, logistical complexity, and effectiveness in infrastructure design; and preventing unintended water subsidies. Response variables in forest PMEs were organized into three broad tiers reflecting increasing complexity and resource intensiveness, with the first tier representing a recommended core set of common measurements. Differences in site conditions combined with unique research questions of experimentalists necessitate careful adaptation of guidelines for forest PMEs to balance local objectives with coordination among experiments. We advocate adoption of a common framework for coordinating forest PME design to enhance cross‐site comparability and advance fundamental knowledge about the response and sensitivity of diverse forest ecosystems to precipitation extremes.

How ecologists define drought, and why we should do better

“…Terrestrial drought impacts are generally studied more in ecosystems with short‐statured vegetation (e.g., grasslands, deserts, savannas) than tall‐statured vegetation (e.g., forests) and this is particularly true for experimental drought studies (Figure inset). This emphasis on shorter systems is likely due in part to the logistical and cost constraints of deploying such experiments in forests (Asbjornsen et al, ; Wullschleger & Hanson, ). Unequal representation among biomes in the ecological drought literature underscores the need for more diverse drought studies, particularly in systems that have been historically understudied.…”

Drought Differentially Affects Growth, Transpiration, and Water Use Efficiency of Mixed and Monospecific Planted Forests

“…Understanding how species interactions and planting design (i.e., mixtures vs. monocultures) impact tradeoffs of productivity, transpiration, and water use efficiency is critical to managing plantations to effectively achieve restoration goals. Throughfall manipulation experiments are becoming a common way of testing hypotheses related to tree responses during drought [12]. While these experiments are important, and can simulate soil moisture stress resulting from less rainfall, a common criticism of this approach is the inability to reproduce the forecasted changes in vapor pressure deficit (VPD) and radiation, both of which have been shown to differ from normal conditions during severe droughts [12].…”

“…Throughfall manipulation experiments are becoming a common way of testing hypotheses related to tree responses during drought [12]. While these experiments are important, and can simulate soil moisture stress resulting from less rainfall, a common criticism of this approach is the inability to reproduce the forecasted changes in vapor pressure deficit (VPD) and radiation, both of which have been shown to differ from normal conditions during severe droughts [12]. Since VPD and radiation have direct effects on tree water use [13][14][15] and because dynamics of radiation and VPD change throughout a drought year, studies of tree drought responses need to take advantage of non-manipulative droughts where meteorological conditions are changing in addition to soil moisture and precipitation changes.…”

Short-term effects of drought on tropical forest do not fully predict impacts of repeated or long-term drought: gas exchange versus growth

“…Here we use the leverage of the only decadal-scale rainfall exclusion experiment in tropical forest to test whether observations made over minutes to months can be used to predict changes in ecological function at longer timescales. Field-based drought experiments are implemented by deflecting away from the soil a fraction of the rainfall that penetrates the canopy (‘throughfall exclusion’, TFE), thus increasing soil moisture deficit [ 28 , 46 ]. This manipulation separates the influences of soil and atmospheric drought on vegetation processes, but it also results in less extreme stress than a natural drought would exert with similarly low rainfall, as the maxima of air temperature and vapour pressure deficit are smaller.…”