Adrian Israel Silva Cardoza

Soy ingeniero en restauración forestal, especialista en manejo del fuego con experiencia en percepción remota y sistemas de información geográfica aplicados al análisis de incendios forestales. Egresado de la Universidad Autónoma Chapingo (2016), obtuvo la maestría en Geomática Aplicada a los Recursos Forestales y Ambientales por la Universidad Juárez del Estado de Durango (2021) y actualmente cursa el Doctorado en Ciencias Geoespaciales en CentroGeo.

Mi trayectoria incluye la colaboración proyectos de investigación con instituciones nacionales e internacionales en el desarrollo de metodologías para la evaluación de severidad post-incendio, la actualización de mapas de áreas prioritarias y la generación de insumos geoespaciales para la toma de decisiones. He liderado proyectos de consultoría para organismos como CONAFOR, CONANP, CONABIO, el USDA Forest Service, la Universidad Autónoma Chapingo y BIOMASA A.C., destacándose por su capacidad para analizar, procesar e interpretar imágenes satelitales y datos de campo. También se ha participado como instructor en talleres y cursos especializados en manejo del fuego, SIG y percepción remota, presentando diversos trabajos de investigación en congresos nacionales e internacionales. Estas contribuciones han sido publicadas en revistas científicas y reconocidas por su rigor técnico y aplicabilidad operativa. Soy miembro activo de la Association for Fire Ecology, la International Association of Wildland Fire, la Red Mexicana de Manejo del Fuego y la Asociación Mexicana de Incendios Forestales y Manejo del Fuego. Su labor se distingue por la responsabilidad, creatividad y compromiso con el fortalecimiento de las capacidades técnicas para el manejo integral del fuego en México y Latinoamérica.


Sessions

11-12
11:15
15min
Integration of field methodology and mobile technology for fire severity assessment in Mexico: The SEVisual MX application
Adrian Israel Silva Cardoza, Yair Uriel Ricardez Diaz, Daniela Moctezuma, José Luis Silván Cárdenas

Background. Post-fire severity assessment is a fundamental component for understanding a wide variety of ecological processes, including greenhouse gas emissions, post-fire soil erosion, tree mortality, and ecosystem resilience. In Mexico where forest fires are recurrent across its temperate, tropical, and semi-arid ecosystems, evaluating the ecological impacts by fire has historically been constrained by the absence of standardized, scalable protocols that effectively link field observations with satellite‑derived severity metrics. Traditional methods, such as the Composite Burn Index and spectral indices like the Relativized Normalized Burn Ratio (RBR), have proven useful globally but require ecosystem‑specific adjustments to ensure accuracy. National efforts to formalize fire severity assessment have produced detailed field methodologies; however, these protocols often demand time, specialized personnel, training, and equipment, limiting their operational adoption by field brigades. This gap between rigorous scientific methods and the need for practical field assessments underscores the necessity for tools that simplify data collection while preserving methodological integrity. In response, the SEVisual MX mobile application was developed to operationalize Mexico’s standardized field severity methodology, enabling efficient, consistent, and georeferenced data collection that supports both ecological research and fire management operations. The study aims to: (1) demonstrate how the app integrates the conceptual and methodological foundations of the Mexican field protocol; (2) illustrate its capacity to standardize and expedite post‑fire severity assessments; and (3) highlight its utility for generating high‑quality field data that can be directly integrated into remote sensing workflows and national fire management systems.
Materials and Methods. The foundational field methodology employs a stratified sampling design, recommending a minimum of 30 plots (15 m radius) per fire, distributed across homogenous patches identified via pre-fire and post-fire spectral indices such as Sentinel-2 derived RBRc. The methodology evaluates two primary components: Vegetation is stratified into tree canopy (DBH > 7.5 cm), understory (DBH < 7.5 cm), regeneration post-fie, and herbaceous layers, while the soil component assesses mineral soil, litter, fermentation, and ash characteristics. Key variables include the proportion of canopy scorch categorized by color (green/unburned, brown/scorched, black/calcined) and the terminal diameter of the finest branches as well as structural changes in organic horizons. SEVisual MX was designed to replicate this methodological structure through a guided digital workflow consisting of three sections: (1) general site information, (2) strata‑specific visual assessments, and (3) automated results. The app captures GPS coordinates directly from the mobile device, functioning fully offline to support remote field operations. Users can classify plots as wildfire or prescribed burn, and as burned or unburned controls. Each stratum includes standardized categorical inputs for coverage, height, scorch proportions, and structural changes, replacing complex measurement tools with guided visual estimation. To ensure data quality, the app requires in-situ georeferenced photographic evidence for each evaluated layer. Once the assessment is completed, SEVisual MX automatically compiles the information into a CSV file accompanied by organized photographic folders, enabling seamless integration into GIS platforms and remote sensing calibration workflows.
Results. The implementation of SEVisual MX demonstrates substantial operational advantages. First, the app reduces the need for large, highly trained brigades; a single technician can complete a full severity assessment using only a mobile device. Second, the app preserves strict conceptual alignment with the original methodology, ensuring that critical ecological indicators—such as canopy scorch, a key predictor of post‑fire pine mortality and plague susceptibility—are captured consistently. The standardized visual categories for soil severity, based on organic matter consumption and structural changes, minimize observer bias and improve reproducibility across teams and ecosystems.
Preliminary field validation indicates that SEVisual MX successfully generates an executive summary and preliminary severity indices for canopy, understory, and soil strata. The exported CSV files provide structured, high‑quality ground‑truth data essential for calibrating medium‑resolution satellite imagery (Sentinel‑2, Landsat) and high‑resolution imagery from Remotely Piloted Aircraft Systems (RPAS). The app’s ability to systematically organize georeferenced photographs enhances post‑processing workflows, enabling rapid verification of field observations and supporting the development of spectral thresholds for severity classification. Early comparisons show strong consistency between app‑based visual assessments and traditional field measurements, confirming the tool’s reliability for operational use.
Conclusions. The integration of a rigorous ecological sampling protocol into the SEVisual MX mobile application represents a significant advancement for fire severity assessment in Mexico. By translating complex variables—such as strata‑specific canopy scorch and soil organic layer consumption—into an intuitive digital interface, the app bridges the longstanding gap between scientific research and field operations. Its standardized data collection framework supports post‑fire ecological restoration planning, hydrological risk mitigation, and decision‑making within integrated fire management strategies. Beyond its methodological contributions, SEVisual MX aligns with national efforts to strengthen institutional fire management capacities. The application is designed to integrate seamlessly with the Sistema de Predicción de Peligro de Incendios Forestales (SPPIF), ensuring that field‑collected severity data directly support spatial analysis, remote sensing calibration, and operational decision‑making. This technological integration reinforces the SPPIF’s ability to characterize fire effects in near‑real time and contributes to the consolidation of a unified national protocol for post‑fire assessment. Moreover, the standardized, reproducible nature of SEVisual MX positions it as a transferable methodology for other Latin American countries, where diverse ecosystems and recurrent fires demand practical, scientifically grounded tools. This app not only advances fire science and management in Mexico but also contributes to the broader development of a Latin American framework for post‑fire severity assessment, fostering comparability, adaptive management, and ecological resilience across the region.

Geographic Information Systems and Cartography
Video Room
11-13
11:15
15min
Dynamic forest fire hazard mapping with artificial intelligence: An integrated geospatial framework for the Trans‑Mexican Volcanic Belt
Adrian Israel Silva Cardoza, Daniela Moctezuma, Hugo Carlos-Martínez

Background
Forest fires are a central ecological process in many regions of Mexico, particularly in temperate and montane ecosystems where fire has historically shaped vegetation structure and regeneration dynamics. However, under current conditions of climate change, land‑use transformation, and expanding human pressure, fire occurrence increasingly leads to high‑risk scenarios for communities, ecosystems, and Protected Natural Areas (ANP). In the central region of the country, decades of fire‑exclusion policies have contributed to fuel accumulation, increasing the likelihood of severe events. At the same time, traditional fire‑use practices, regulatory constraints, and heterogeneous patterns of land occupation create spatially differentiated ignition conditions. Existing monitoring systems often rely on static or coarse‑resolution approaches that do not adequately represent the spatial variability of fuels, microclimates, or anthropogenic drivers. These limitations underscore the need for methodologies that integrate meteorological, fuels, anthropogenic, and topographic predictors within a machine‑learning framework capable of capturing the dynamic conditions that influence fire occurrence probability. Combining these components enables the construction of hazard maps that anticipate critical scenarios and support adaptive fire‑management strategies. This work was carried out with the support of the Mexican Fund for the Conservation of Nature (FMCN) and the Development Bank KfW, whose collaboration has strengthened ongoing efforts to advance geospatial tools for fire‑risk assessment and sustainable landscape management.
Objective
Develop and implement a methodology for generating forest fire hazard maps in the Central Region and Trans‑Mexican Volcanic Belt. The aim is to explain spatial patterns of fire hazard by evaluating methodological improvements, including the integration of multivariate machine‑learning algorithms, optimized predictor selection, and threshold calibration for operational decision‑making.
Materials and methods
The study area covers 164,203 km² across 13 states and 771 municipalities. It includes 19 federal ANP that represent key conservation areas. The region exhibits strong altitudinal gradients, diverse vegetation types, and marked climatic variability, all of which influence fuel availability and ignition patterns. Fire occurrence records from 2012–2024 were obtained from the national forest agency. A binary presence–absence dataset was constructed by filtering events within forested areas and generating pseudo‑absences outside a 1‑km exclusion buffer around each ignition point. This approach reduces spatial autocorrelation and ensures that absence samples represent distinct environmental conditions. The final dataset includes annual ignition points and corresponding absence samples, each associated with vegetation type and year of occurrence. The independent variables were organized into four thematic components. The anthropogenic component included distance to protected areas, settlements, and agricultural zones, as well as population density derived, representing human activities that drive most ignitions. The forest‑fuel component incorporated spectral indices from VIIRS imagery—such as GOSAVI, NBR, and NDVI—processed into 60‑day composites to reduce atmospheric noise and capture short‑term vegetation conditions. The topographic component included slope, terrain roughness, and related metrics derived from the national elevation model and resampled to 1 km to ensure spatial consistency. The meteorological component used Daymet reanalysis to compute 60‑day composites of vapor pressure deficit, accumulated precipitation, accumulated radiation, and temperature extremes, representing atmospheric conditions that influence fuel moisture and ignition likelihood.
The dataset was split into training (60%), validation (20%), and external testing (20%) subsets using stratified sampling to preserve class proportions. Variables were normalized to the [0,1] range when required by the algorithms. A multi‑step procedure was applied: (i) Univariate AUC analysis to identify predictors with the highest discriminative capacity, (ii) Collinearity assessment using Pearson correlation and variance inflation factors, (iii) Stability verification using Y‑Scrambling to ensure that predictive relationships were not random, (iv) Four predictors were selected for the final model: distance to ANP (adan), terrain roughness (teri), 60‑day vapor pressure water (vp), and spectral index gosavi.
Five algorithms were evaluated: Random Forest, XGBoost, Support Vector Machines with radial kernel, k‑Nearest Neighbors, and logistic regression. Models were trained using stratified 10‑fold cross‑validation. Performance metrics included AUC, accuracy, sensitivity, specificity, Matthews correlation coefficient, and F1‑score. The two best‑performing models (Random Forest and XGBoost) were optimized using grid search. Final evaluation was conducted on the external test set, including threshold optimization using the Youden index. The final model was applied to a multiband raster stack of predictors aligned in resolution and extent. Continuous probabilities were classified into five hazard levels. Spatial residuals were computed to identify areas of over‑ or under‑estimation.
Results and discussion
Random Forest achieved the highest performance, with an AUC of 0.906, followed by XGBoost with 0.895. Both models showed consistent behavior across folds, indicating stable generalization. Threshold optimization increased sensitivity from 0.720 to 0.807 for Random Forest, improving detection of ignition events with a moderate reduction in specificity. The selected predictors captured key spatial gradients: ignitions were more frequent within 50 km of ANP, in areas of intermediate terrain roughness, under low atmospheric moisture (vapor pressure near 500 Pa), and in vegetation with intermediate values, reflecting moderate biomass and canopy openness. The results highlight the interaction between climatic drivers, fuel conditions, and human activities. The inclusion of distance to ANP captured spatial gradients associated with land‑use restrictions and community fire‑use practices. The 60‑day meteorological composites effectively represented short‑term atmospheric conditions relevant to ignition. Limitations include positional uncertainty in fire records, coarse spatial resolution of Daymet, and the absence of high‑resolution wind and humidity data. Future work should integrate hot-spots, thermal anomalies, and alternative climate datasets such as CHIRPS or ERA5‑Land, as well as medium‑resolution spectral data from Sentinel‑2 or Landsat.
Conclusion
The study demonstrates the feasibility of integrating machine learning with meteorological, spectral, and topographic predictors to generate dynamic fire hazard maps at regional scale. The methodology is scalable and suitable for operational implementation. Future efforts should evaluate multiple temporal windows, incorporate satellite‑derived ignition proxies, and expand validation across additional regions.

Artificial Intelligence and Geographical Information Science
Video Room