Bedrock mediates responses of ecosystem productivity to climate variability

Climate change significantly affects Earth's ecosystems. However, current understanding primarily focuses on mean climate states, neglecting the role of climate variability in ecosystem structure and functioning. Ecosystem responses to climatic variability are crucial for understanding ecosystem resilience. This study explores the influence of bedrock lithology (and associated soil and regolith properties like thickness, porosity, and permeability) on the sensitivity of terrestrial ecosystem primary productivity to changes in climate water deficit (CWD), a measure of hydroclimatic variability. Bedrock lithology affects productivity by influencing plant-available water storage. While bedrock typically has low porosity and conductivity, differences in mineral weathering rates lead to varying regolith formation, permeability, and nutrient production. For example, carbonate rocks' high solubility creates preferential flow paths, increasing permeability and reducing water retention. Regolith thickness significantly impacts water storage, especially benefiting deep-rooted plants during droughts. Although bedrock and regolith properties affect productivity, their impact on ecosystem sensitivity to climatic variability is less understood. Ecosystem sensitivity can be negative (productivity decreases with increased CWD) or positive (productivity increases with increased CWD), depending on primary constraints (water, temperature). This study addresses three knowledge gaps: quantifying lithologic effects on ecosystem sensitivity globally, using a global framework to evaluate bedrock lithology's effects, and isolating the underlying mechanisms by considering mediating variables like regolith thickness, permeability, and porosity. Newly available global datasets allow for a comprehensive analysis of these effects.

Previous research has shown that bedrock lithology influences ecosystem productivity through variations in plant-available water storage and nutrient release. Studies in karst regions demonstrate significantly higher ecosystem sensitivity to interannual climate variations due to lower water-holding capacity in karst bedrock. The importance of regolith water storage for deep-rooted plants during droughts has also been highlighted. While these studies have established links between bedrock geochemistry and ecosystem productivity, the mediating role of regolith properties and the global scale patterns have not been fully investigated due to limitations in data availability. This study utilizes newly available global datasets to address these gaps, providing a more comprehensive understanding of bedrock's role in ecosystem sensitivity.

The study used a two-stage approach. First, it quantified global patterns of ecosystem sensitivity to climate variability using gridded global measurements of CWD and ecosystem productivity (inferred from GIMMS NDVI data). Ecosystem sensitivity was defined as the change in NDVI relative to the change in CWD. A Bayesian linear regression was used for each grid cell to determine the relationship between CWD and NDVI. The slopes of these regressions represent ecosystem sensitivity. Second, it used high-resolution data on lithology and hydro-lithological characteristics to evaluate the effects of bedrock lithology on ecosystem sensitivity. A weighted Bayesian hierarchical regression model was used to explain the variation in ecosystem sensitivity across grid cells, weighting each sensitivity value according to its variance. This model included bedrock lithology, regolith porosity, regolith permeability, soil thickness, and regolith thickness as explanatory variables, and accounted for spatial autocorrelation among grid cells. The model was structured to estimate region-specific effects, with regions categorized based on the sign of ecosystem sensitivity and primary productivity constraints (energy-limited, water-limited, humid tropical, hyper-arid). The study also investigated model robustness using alternative datasets for ecosystem productivity (GPP) and sensitivity calculations focusing on growing seasons. The study assessed model fit using the coefficient of determination (R2) from a regression of observed versus predicted sensitivity.

Globally, the mean sensitivity of ecosystem productivity to CWD was low, with most ecosystems (63% of grid cells) exhibiting negative sensitivity, where increased CWD reduced productivity. A significant portion (45%) of these negatively sensitive grids showed statistical significance. Positive sensitivity was observed in 37% of ecosystems, with significant positive sensitivity in 13.3% of grids. Positive sensitivity was most prominent in energy-limited (high-latitude, high-altitude) and humid tropical regions, while hyper-arid regions showed a smaller positive response. In negatively sensitive regions, higher regolith porosity and permeability increased sensitivity to climate variability. In positively sensitive regions, these properties reduced sensitivity. In energy-limited regions, increased regolith porosity and permeability reduced the positive response to warming. Increased soil and regolith thickness reduced sensitivity in negatively sensitive regions, and increased sensitivity in positively sensitive regions. After accounting for the effects of soil and regolith properties, significant residual effects of bedrock lithology were found in positively sensitive regions. Sedimentary rocks generally dampened the positive response to CWD in energy-limited and hyper-arid regions, while plutonic rocks enhanced it. In humid tropical regions, sedimentary rocks increased sensitivity. The effect sizes of changes in regolith porosity, permeability, soil thickness and regolith thickness on ecosystem sensitivity ranged from 3% to 34% of the region-specific mean. The study also explored alternative datasets and growing season data for robustness checks and found that major conclusions remained consistent.

The study's findings reveal a complex relationship between bedrock lithology, regolith properties, and ecosystem sensitivity to climate variability. The diverse responses highlight the importance of considering regional context and primary productivity constraints. Regolith permeability and porosity influence water-holding capacity, directly impacting ecosystem response to CWD. Increased permeability decreases water-holding capacity, making ecosystems more vulnerable to CWD. In water-limited regions (negative sensitivity), thick regolith and soil buffers against CWD effects. In energy-limited regions, thicker regolith and soil may mitigate water limitations during warming. The positive sensitivity in humid tropical regions reflects productivity increases with reduced precipitation, likely due to decreased oxygen limitation and reduced nutrient leaching. Hyper-arid regions' positive sensitivity may be attributed to CO2 fertilization, and here thicker regolith and soil enhanced this effect. Residual effects of bedrock lithology, particularly sedimentary rocks, suggest additional controls beyond water-holding capacity, potentially involving nutrient availability and plant adaptation to hydrological variability. This complex interplay between lithology, regolith properties, and climate impacts ecosystem resilience.

This study demonstrates the significant influence of bedrock and its weathering products on global ecosystem sensitivity to climate variability, primarily through their control of water-holding capacity. Regolith properties and soil thickness substantially modify ecosystem responses to climate change. However, regional context, primarily limited by water or energy availability, and bedrock lithology independently modify these relationships. Future research could explore the interplay of these factors with other ecosystem attributes (biodiversity, topography) to provide a more comprehensive understanding of ecosystem resilience.

The study's reliance on existing datasets introduces limitations. The accuracy of the NDVI and CWD datasets might influence the results. The study did not explicitly correct for agricultural areas, which could affect sensitivity estimates. While robustness checks were performed, other confounding factors could influence ecosystem sensitivity. Causal inferences are based on statistical correlations and require further investigation.