Earth & Environmental Sciences

Biogeography is a dynamic field that has transformed dramatically over the last few decades from being necessarily descriptive to become a rigorous science. Major recent areas of growth have included phylogenetics and phylogeography, microbial biogeography and metagenomics, and macroecology. However, the welcome recent deluge of massive amounts of data, in particular from genomics, museum specimens, and field observations, as well as environmental information, is posing a huge challenge to the field. The society has several key roles, not only to serve as a home for researchers in the field and enabling interaction among them, but also: (1) to provide a forum to facilitate awareness and use of rapidly developing tools and data; (2) to encourage a solid foundation in organismal research, with emphasis on field and museum based resources; (3) to promote global connections; and (4) to cultivate interdisciplinarity, such that the predictive capabilities of the field can be used to inform management and policy.

Russias invasion of Ukraine continues to have a devastating effect on the well-being of Ukrainians and their environment. We evaluated a major environmental hazard caused by the war: the potential for groundwater contamination in proximity to the Zaporizhzhia Nuclear Power Plant (NPP). We quantified groundwater vulnerability with the DRASTIC index, which was originally developed by the United States Environmental Protection Agency and has been used at various locations worldwide to assess relative pollution potential. We found that there are two major gradients of groundwater vulnerability in the region: (1) broadly higher risk to the northeast of the NPP and lower risk to the southeast driven by a regional gradient in water availability and water table depth; and (2) higher risk in proximity to the channels and floodplains of the Dnipro River and tributaries, which host coarser-textured soils and sedimentary deposits. We also found that the DRASTIC vulnerability index can be used to identify and prioritize groundwater well-network monitoring. These and more detailed assessments will be necessary to prioritize monitoring and remediation strategies across Ukraine in the event of a nuclear accident, and more broadly demonstrate the utility of the DRASTIC approach for prognostic contamination risk assessment.

Undocumented Orphaned Wells (UOWs) are wells without an operator that have limited or no documentation with regulatory authorities. An estimated 310,000 to 800,000 UOWs exist in the United States (US), whose locations are largely unknown. These wells can potentially leak methane and other volatile organic compounds to the atmosphere, and contaminate groundwater. In this study, we developed a novel framework utilizing a state-of-the-art computer vision neural network model to identify the precise locations of potential UOWs. The U-Net model is trained to detect oil and gas well symbols in georeferenced historical topographic maps, and potential UOWs are identified as symbols that are further than 100 m from any documented well. A custom tool was developed to rapidly validate the potential UOW locations. We applied this framework to four counties in California and Oklahoma, leading to the discovery of 1301 potential UOWs across >40,000 km². We confirmed the presence of 29 UOWs from satellite images and 15 UOWs from magnetic surveys in the field with a spatial accuracy on the order of 10 m. This framework can be scaled to identify potential UOWs across the US since the historical maps are available for the entire nation.

Machine learning (ML) applications in soil science have significantly increased over the past two decades, reflecting a growing trend towards data-driven research addressing soil security. This extensive application has mainly focused on enhancing predictions of soil properties, particularly soil organic carbon, and improving the accuracy of digital soil mapping (DSM). Despite these advancements, the application of ML in soil science faces challenges related to data scarcity and the interpretability of ML models. There is a need for a shift towards Soil Science-Informed ML (SoilML) models that use the power of ML but also incorporate soil science knowledge in the training process to make predictions more reliable and generalisable. This paper proposes methodologies for embedding ML models with soil science knowledge to overcome current limitations. Incorporating soil science knowledge into ML models involves using observational priors to enhance training datasets, designing model structures which reflect soil science principles, and supervising model training with soil science-informed loss functions. The informed loss functions include observational constraints, coherency rules such as regularisation to avoid overfitting, and prior or soil-knowledge constraints that incorporate existing information about the parameters or outputs. By way of illustration, we present examples from four fields: digital soil mapping, soil spectroscopy, pedotransfer functions, and dynamic soil property models. We discuss the potential to integrate process-based models for improved prediction, the use of physics-informed neural networks, limitations, and the issue of overparametrisation. These approaches improve the relevance of ML predictions in soil science and enhance the models’ ability to generalise across different scenarios while maintaining soil science principles, transparency and reliability.

Statistical modeling of high-dimensional matrix-valued data motivates the use of a low-rank representation that simultaneously summarizes key characteristics of the data and enables dimension reduction. Low-rank representations commonly factor the original data into the product of orthonormal basis functions and weights, where each basis function represents an independent feature of the data. However, the basis functions in these factorizations are typically computed using algorithmic methods that cannot quantify uncertainty or account for basis function correlation structure a priori. While there exist Bayesian methods that allow for a common correlation structure across basis functions, empirical examples motivate the need for basis function-specific dependence structure. We propose a prior distribution for orthonormal matrices that can explicitly model basis function-specific structure. The prior is used within a general probabilistic model for singular value decomposition to conduct posterior inference on the basis functions while accounting for measurement error and fixed effects. We discuss how the prior specification can be used for various scenarios and demonstrate favorable model properties through synthetic data examples. Finally, we apply our method to two-meter air temperature data from the Pacific Northwest, enhancing our understanding of the Earth system's internal variability.

Future climate presents conflicting implications for forest biomass. We evaluate how plant hydraulic traits, elevated CO₂ levels, warming, and changes in precipitation affect forest primary productivity, evapotranspiration, and the risk of hydraulic failure. We used a dynamic vegetation model with plant hydrodynamics (FATES-HYDRO) to simulate the stand-level responses to future climate changes in a wet tropical forest in Barro Colorado Island, Panama. We calibrated the model by selecting plant trait assemblages that performed well against observations. These assemblages were run with temperature and precipitation changes for two greenhouse gas emission scenarios (2086-2100: SSP2-45, SSP5-85) and two CO₂ levels (contemporary, anticipated). The risk of hydraulic failure is projected to increase from a contemporary rate of 5.7% to 10.1-11.3% under future climate scenarios, and, crucially, elevated CO₂ provided only slight amelioration. By contrast, elevated CO₂ mitigated GPP reductions. We attribute a greater variation in hydraulic failure risk to trait assemblages than to either CO₂ or climate. Our results project forests with both faster growth (through productivity increases) and higher mortality rates (through increasing rates of hydraulic failure) in the neo-tropics accompanied by certain trait plant assemblages becoming nonviable.

Lianas (woody climbers) are crucial components of tropical forests and they have been increasingly recognized to have profound effects on tropical forest carbon dynamics. Despite their importance, lianas' representation in vegetation models remains limited, partly due to the complexity of liana-tree dynamics and the diversity in liana life history strategies. This paper provides a comprehensive review of advances and challenges for mechanistically representing lianas in forest ecosystem models and a proposed path towards effectively representing lianas in these models. Defining a liana plant functional type is a significant challenge because of the high morphological and physiological diversity amongst liana species, and because of their structural association with trees. Here, we identify critical liana traits that likely should contribute to establishing a liana plant functional type, along with key processes to properly represent lianas in ecosystem models. Subsequently, we discuss a variety of possible liana implementation strategies with their associated strengths, limitations, computational costs and data requirements. A fundamental redesign of the tree-centric demographic vegetation models seems appropriate to accommodate the unique growth and competition strategies of lianas. We illustrate the potential of such models with a single-site case study where we disentangle putative mechanisms of liana increasing abundance. Furthermore, we underscore the critical need for comprehensive liana demographic and functional data (including long-term, physiological, and pantropical observations) for the qualitative implementation and evaluation in the proposed modeling efforts. Currently, there is a scarcity of liana data and the data that do exist have a neotropical bias. We finally introduce a new liana functional trait database that can centralize existing liana trait data, incentivize improved data gathering and thus facilitate model development and scientific analyses.

Convective storms with strong downdrafts create windthrows: snapped and uprooted trees that locally alter the structure, composition, and carbon balance of forests. Comparing Landsat imagery from subsequent years, we documented temporal and spatial variation in the occurrence of large (≥30 ha) windthrows across the Amazon basin from 1985 to 2020. Over 33 individual years, we detected 3179 large windthrows. Windthrow density was greatest in the central and western Amazon regions, with ∼33% of all events occurring in ∼3% of the monitored area. Return intervals for large windthrows in the same location of these “hotspot” regions are centuries to millennia, while over the rest of the Amazon they are >10,000 years. Our data demonstrate a nearly 4-fold increase in windthrow number and affected area between 1985 (78 windthrows and 6,900 ha) and 2020 (264 events and 32,170 ha), with more events of >500 ha size since 1990. Such extremely large events (>500 ha up to 2,543 ha) are responsible for interannual variation in the overall median (84 ± 5.2 ha; ±95% CI) and mean (147 ± 13 ha) windthrow area, but we did not find significant temporal trends in the size distribution of windthrows with time. Our results document increased damage from convective storms over the past 40 years in the Amazon, filling a gap in temporal records for tropical regions. Our publicly accessible large windthrow database provides a valuable tool for exploring dynamic conditions leading to damaging storms and their ecological impact on Amazon forests.