Search for: All records

Abstract Purpose of ReviewArtificial intelligence (AI) is disrupting science and discovery across disciplines, offering new modes of inquiry that are changing how questions are asked and answered and upsetting established norms. In this paper, we review the state of the art of AI in landscape ecology and offer six areas of opportunity for landscape ecologists to capitalize on AI tools moving forward. These areas include geospatial AI (GeoAI), geometric AI, Explainable AI (xAI), generative AI (GenAI), Natural Language Processing (NLP), and robotics. Recent FindingsLandscape ecology has a long history of using AI, notably machine learning methods for image classification tasks, agent-based modeling, and species distribution modeling but also knowledge representation and automated reasoning for landscape generation and spatial planning. Methods have become more diverse and complex in recent years, with a new generation of AI-based tools rapidly emerging. These new tools have potential to improve how landscape ecologists map, measure, and model landscape patterns and processes as well as improve the explainability of model outputs. SummaryThere are many untapped opportunities for landscape ecologists to leverage emerging AI-based tools in research and practice including generating virtual landscapes for simulating processes such as wildfires and leveraging natural language processing to generate new insights from text data. Regardless of the application, researchers using AI tools must also consider the ethical implications of data and algorithmic biases and critically assess how these methods can be used responsibly. 

Abstract Droughts are a natural hazard of growing concern as they are projected to increase in frequency and severity for many regions of the world. The identification of droughts and their future characteristics is essential to building an understanding of the geography and magnitude of potential drought change trajectories, which in turn is critical information to manage drought resilience across multiple sectors and disciplines. Adding to this effort, we developed a dataset of global historical and projected future drought indices over the 1980–2100 period based on downscaled CMIP6 models across multiple shared socioeconomic pathways (SSP). The dataset is composed of two indices: the Standardized Precipitation Index (SPI) and Standardized Precipitation Evapotranspiration Index (SPEI) for 23 downscaled global climate models (GCMs) (0.25-degree resolution), including historical (1980–2014) and future projections (2015–2100) under four climate scenarios: SSP1-2.6, SSP2-4.5, SSP3-7.0, and SSP5-8.5. The drought indices were calculated for 3-, 6- and 12-month accumulation timescales and are available as gridded spatial datasets in a regular latitude-longitude format at monthly time resolution. 

Mapping agricultural fields using high-resolution satellite imagery and deep learning (DL) models has advanced significantly, even in regions with small, irregularly shaped fields. However, effective DL models often require large, expensive labeled datasets, which are typically limited to specific years or regions. This restricts the ability to create annual maps needed for agricultural monitoring, as changes in farming practices and environmental conditions cause domain shifts between years and locations. To address this, we focused on improving model generalization without relying on yearly labels through a holistic approach that integrates several techniques, including an area-based loss function, Tversky-focal loss (TFL), data augmentation, and the use of regularization techniques like dropout. Photometric augmentations helped encode invariance to brightness changes but also increased the incidence of false positives. The best results were achieved by combining photometric augmentation, TFL, and Monte Carlo dropout, although dropout alone led to more false negatives. Input normalization also played a key role, with the best results obtained when normalization statistics were calculated locally (per chip) across all bands. Our U-Net-based workflow successfully generated multi-year crop maps over large areas, outperforming the base model without photometric augmentation or MC-dropout by 17 IoU points. 

Explaining deep learning models operating on time series data is crucial in various applications of interest which require interpretable and transparent insights from time series signals. In this work, we investigate this problem from an information theoretic perspective and show that most existing measures of explainability may suffer from trivial solutions and distributional shift issues. To address these issues, we introduce a simple yet practical objective function for time series explainable learning. The design of the objective function builds upon the principle of information bottleneck (IB), and modifies the IB objective function to avoid trivial solutions and distributional shift issues. We further present TimeX++, a novel explanation framework that leverages a parametric network to produce explanation-embedded instances that are both in-distributed and label-preserving. We evaluate TimeX++ on both synthetic and real-world datasets comparing its performance against leading baselines, and validate its practical efficacy through case studies in a real-world environmental application. Quantitative and qualitative evaluations show that TimeX++ outperforms baselines across all datasets, demonstrating a substantial improvement in explanation quality for time series data.