Geospatial code generation is emerging as a key direction in the integration of artificial intelligence and geoscientific analysis. However, there remains a lack of standardized tools for automatic evaluation in this domain. To address this gap, we propose AutoGEEval, the first multimodal, unit-level automated evaluation framework for geospatial code generation tasks on the Google Earth Engine (GEE) platform powered by large language models (LLMs). Built upon the GEE Python API, AutoGEEval establishes a benchmark suite (AutoGEEval-Bench) comprising 1325 test cases that span 26 GEE data types. The framework integrates both question generation and answer verification components to enable an end-to-end automated evaluation pipeline—from function invocation to execution validation. AutoGEEval supports multidimensional quantitative analysis of model outputs in terms of accuracy, resource consumption, execution efficiency, and error types. We evaluate 18 state-of-the-art LLMs—including general-purpose, reasoning-augmented, code-centric, and geoscience-specialized models—revealing their performance characteristics and potential optimization pathways in GEE code generation. This work provides a unified protocol and foundational resource for the development and assessment of geospatial code generation models, advancing the frontier of automated natural language to domain-specific code translation.

Schools across the United States and around the world canceled in-person classes beginning in March 2020 to contain the spread of the COVID-19 virus, a public health emergency. Many empirical pieces of research have demonstrated that educational institutions aid students’ overall growth and studies have stressed the importance of prioritizing in-person learning to cultivate social values through education. Two years into the COVID-19 pandemic, policymakers and school administrators have been making plans to reopen schools. However, few scientific studies had been done to support planning classroom seating while complying with the social distancing policy. To ensure a safe return to campus, we designed a ‘community-safe’ method for classroom management that incorporates social distancing and computes seating capacity. In this paper, we present custom GIS tools developed for two types of classroom settings – classrooms with fixed seating and classrooms with movable seating. The fixed model tool is based on an optimized backtracking algorithm. Our flexible model tool can consider various classroom dimensions, fixtures, and a safe social distance. The tool is built on a python script that can be executed to calculate revised seating capacity to maintain a safe social distance for any defined space. We present a real-world implementation of the system at Eastern Michigan University, United States, where it was used to support campus reopening planning in 2020. Our proposed GIS-based technique could be applicable for seating planning in other indoor and outdoor settings.

The service network is capable of addressing large-scale service composition. However, existing service network works still have several limitations. Prior knowledge, such as expert-defined service chains, is not incorporated into the service network. QoS constraints are less considered in the service network, and thus the generated service chain does not always satisfy the optimal QoS constraints. Additionally, some basic services also require outputs to be used directly as inputs, which the service network cannot provide. To address these limitations, this paper proposes a geospatial service web (GSW) model named SR-QoS-GSW that incorporates service semantic relationships and QoS information. The SR-QoS-GSW model consists of atomic services and composite services that consider QoS, processing services, data services, and relationships among them. A SR-QoS-GSW prototype was developed using 570 atomic services and 27 composite services and evaluated using two case studies—a river network extraction and an urban housing selection. Then, the information entropy and time complexity between SR-QoS-GSW and the existing service network were compared. The results show that geospatial service chains can be created more efficiently by incorporating existing service chains as composite services. Integrating QoS information into the GSW would allow service composition algorithms to generate service chains that satisfy optimal QoS constraints. The outputs of services used as new inputs with additional self-matching relationships also give the service network greater flexibility. Finally, the analysis of the information entropy and time complexity verified the increased diversity and decreased the search space of the SR-QoS-GSW.

AlphaEarth Foundations (AEF) unifies multi-modal and multi-temporal observations into analysis-ready low-dimensional representations, introducing a representation-driven paradigm that addresses key limitations of traditional task-centric approaches, including high engineering complexity, limited cross-regional transferability, and strong dependence on labeled data. This paper presents the first systematic review of AEF, synthesizing 23 research articles published up to March 2026 from three perspectives: conceptual positioning, technical framework, and application practices. It first clarifies the distinctions between AEF and related paradigms, including foundation models, remote sensing large models, and existing unified embedding approaches. It then summarizes the organization of its data system and key technical components. Based on a systematic literature survey, the paper further provides a structured synthesis of current studies in terms of thematic and regional distribution, scientific questions, usage patterns, and evaluation methods. The analysis indicates that AEF represents an important step toward a paradigm shift from task-driven to representation-driven Earth observation, offering clear advantages in reducing engineering barriers, enabling cross-regional transfer, and establishing a unified environmental semantic foundation. However, limitations remain in temporal resolution, semantic interpretability, and performance in specific task scenarios. Future work should focus on enhancing dynamic representation capabilities, cross-domain adaptation mechanisms, multi-source integration frameworks, systematic evaluation and annotation systems, and asset-oriented applications. By delineating the capability boundaries and applicable contexts of AEF from both methodological and empirical perspectives, this study provides a systematic reference for the standardized development and rational application of unified surface embeddings as an emerging geospatial information infrastructure.

Pixel-level slum mapping remains constrained by limited cross-city generalisation, the lack of continuous estimates of intra-pixel slum coverage, and weak comparability across geographic contexts. AlphaEarth Foundations (AEF), which provides globally consistent 64-dimensional annual surface embeddings at 10 m resolution, offers an analysis-ready representation for scalable and lightweight slum monitoring. However, its applicability to slum mapping remains unclear because slum conditions are only indirectly observable from Earth observation data and are jointly shaped by built-form and socioeconomic processes. We systematically evaluate AEF for binary slum classification and sub-pixel slum cover fraction estimation across 12 cities and 69 city–year pairs from 2017 to 2024, using high-resolution GRAM-derived pseudo-masks as supervisory labels. The evaluation spans four training strategies, two protocols (random split and 3 × 3 spatial block cross-validation), six auxiliary feature configurations, and five baseline models, complemented by representation-level analyses (PCA, SHAP) and full-AOI mapping. Five findings emerge. (1) Same-city cross-year training is optimal under both protocols (median spatial F1 = 0.616, R² = 0.466); temporal expansion outperforms cross-city transfer, indicating city-scale representational drift. (2) Regression R² is driven primarily by zero/non-zero boundary discrimination: positive-pixel R² is consistently negative across all cities, revealing limited capacity to model intra-pixel variation in sub-pixel slum cover fraction at 10 m. (3) PC36 is consistently top-ranked across tasks; classification saturates at k = 32 while regression remains unsaturated at k = 64. (4) POI features yield the largest density gain (ΔR² = +0.064). (5) For six cities meeting dual-task usability thresholds, full-AOI inference across 2017–2024 preserves slum cluster structure (mean SSIM = 0.926). The study delineates the capabilities and complementarity needs of foundation-model embeddings for slum monitoring.