Search for: All records

Drylands are a critical part of the earth system in terms of total area, socioeconomic and ecological importance. However, while drylands are known for their contribution to inter-annual atmospheric CO 2 variability, they are sometimes overlooked in discussions of global carbon stocks. Here, in preparation for the November 2021 UN Climate Change Conference (COP26), we review dryland systems with emphasis on their role in current and future carbon storage, response to climate change and potential to contribute to a carbon neutral future. Current estimates of carbon in dryland soils and vegetation suggest they are significant at global scale, containing approximately 30% of global carbon in above and below-ground biomass, and surface-layer soil carbon (top 30 cm). As ecosystems that are limited by water, the drylands are vulnerable to climate change. Climate change impacts are, however, dependent on future trends in rainfall that include both drying and wetting trends at regional scales. Regional rainfall trends will initiate trends in dryland productivity, vegetation structure and soil carbon storage. However, while management of fire and herbivory can contribute to increased carbon sequestration, impacts are dependent on locally unique ecosystem responses and climate-soil-plant interactions. Similarly, while community based agroforestry initiatives have been successful in some areas, large-scale afforestation programs are logistically infeasible and sometimes ecologically inappropriate at larger scales. As climate changes, top-down prescriptive measures designed to increase carbon storage should be avoided in favour of locally-adapted approaches that balance carbon management priorities with local livelihoods, ecosystem function, biodiversity and cultural, social and economic priorities. 

1. The ecological and environmental science communities have embraced machine learning (ML) for empirical modelling and prediction. However, going beyond prediction to draw insights into underlying functional relationships between response variables and environmental ‘drivers’ is less straightforward. Deriving ecological insights from fitted ML models requires techniques to extract the ‘learning’ hidden in the ML models. 2. We revisit the theoretical background and effectiveness of four approaches for deriving insights from ML: ranking independent variable importance (Gini importance, GI; permutation importance, PI; split importance, SI; and conditional permutation importance, CPI), and two approaches for inference of bivariate functional relationships (partial dependence plots, PDP; and accumulated local effect plots, ALE). We also explore the use of a surrogate model for visualization and interpretation of complex multi-variate relationships between response variables and environmental drivers. We examine the challenges and opportunities for extracting ecological insights with these interpretation approaches. Specifically, we aim to improve interpretation of ML models by investigating how effectiveness relates to (a) interpretation algorithm, (b) sample size and (c) the presence of spurious explanatory variables. 3. We base the analysis on simulations with known underlying functional relationships between response and predictor variables, with added white noise and the presence of correlated but non-influential variables. The results indicate that deriving ecological insight is strongly affected by interpretation algorithm and spurious variables, and moderately impacted by sample size. Removing spurious variables improves interpretation of ML models. Meanwhile, increasing sample size has limited value in the presence of spurious variables, but increasing sample size does improves performance once spurious variables are omitted. Among the four ranking methods, SI is slightly more effective than the other methods in the presence of spurious variables, while GI and SI yield higher accuracy when spurious variables are removed. PDP is more effective in retrieving underlying functional relationships than ALE, but its reliability declines sharply in the presence of spurious variables. Visualization and interpretation of the interactive effects of predictors and the response variable can be enhanced using surrogate models, including three-dimensional visualizations and use of loess planes to represent independent variable effects and interactions. 4. Machine learning analysts should be aware that including correlated independent variables in ML models with no clear causal relationship to response variables can interfere with ecological inference. When ecological inference is important, ML models should be constructed with independent variables that have clear causal effects on response variables. While interpreting ML models for ecological inference remains challenging, we show that careful choice of interpretation methods, exclusion of spurious variables and adequate sample size can provide more and better opportunities to ‘learn from machine learning’.