Karel Aldrin Sánchez Hernández

Páramo ecosystems are among the most biodiverse and hydrologically critical landscapes on Earth, yet their long-term sustainability is increasingly threatened by climate-driven wildfires. Vegetation Cover Fires (VCFs) in these high-altitude environments degrade carbon stocks, disrupt freshwater regulation, and undermine biodiversity conservation goals central to the UN Sustainable Development Goals (SDGs 13, 15, and 6). Between 2001 and 2023, 128 fire events consumed approximately 815 ha in the Sumapaz locality (the world’s largest continuous páramo), representing 64.9% of all fires recorded across Bogotá’s 20 localities. Despite this disproportionate ecological and social impact, no spatially explicit, operational risk management framework has been available for the region, representing a critical sustainability governance gap. This study addresses that gap by proposing an integrated climate-adaptive risk assessment and management strategy based on (i) the Canadian Forest Fire Danger Rating System Fire Weather Index (FWI), derived from ERA5 reanalysis climate data; (ii) spatial and temporal hotspot analysis of MODIS FIRMS active fire detections; and (iii) IDEAM’s multi-component vulnerability and threat scoring protocol. Spatial data were processed using ArcGIS, and FWI sub-indices were computed for each month of the 2001–2023 period. The FWI averaged 0.78 (low danger) across the study period yet peaked at 13.7 in February 2010 (moderate-to-high danger), consistent with the year of highest recorded fire activity (19 events). High- and very high-risk areas (3.70% combined) coincide with slopes >25%, the presence of the invasive and pyrogenic Ulex europaeus, and proximity to populated and agricultural lands. This study concludes with a three-pillar risk management framework—risk knowledge, risk reduction, and disaster management—providing spatially targeted, operationally viable strategies for local and institutional actors that directly support the sustainable conservation of páramo ecosystem services (water supply, carbon sequestration, biodiversity). The framework is designed to be updatable on a monthly basis using freely available ERA5 data, enabling continuous adaptive governance of wildfire risk as a contribution to long-term territorial sustainability. Limitations regarding MODIS detection uncertainty, ERA5 spatial resolution in complex terrain, and the need for probabilistic modeling are explicitly acknowledged.

Vegetation Cover Fires (VCFs) are recurring disturbances that threaten biodiversity, hydrological regulation, and carbon storage. Between 2001 and 2023, 128 fire events consumed approximately 815 ha in Sumapaz, representing 64.9% of all fires recorded across Bogotá's 20 localities. Despite this disproportionate impact, no spatially explicit, operational risk management framework has been available for the region. This study addresses that gap by proposing an integrated risk assessment and management strategy based on (i) the Canadian Forest Fire Danger Rating System Fire Weather Index (FWI), derived from ERA5 reanalysis climate data; (ii) spatial and temporal hot-spot analysis of MODIS FIRMS active-fire detections; and (iii) IDEAM's multi-component Vulnerability and Threat scoring protocol. Spatial data were processed using ArcGIS, and FWI sub-indices were computed for each month of the 2001–2023 period. The FWI averaged 0.78 (low danger) across the study period yet peaked at 13.7 in February 2010 (moderate-to-high danger), consistent with the year of highest recorded fire activity (19 events). High and very high-risk areas (3.70% combined) coincide with slopes >25%, the presence of the invasive and pyrogenic Ulex europaeus, and proximity to populated and agricultural lands. The study concludes with a three-pillar risk management framework risk knowledge, risk reduction, and disaster management providing spatially targeted, operationally viable strategies for local and institutional actors. Limitations regarding MODIS detection uncertainty, ERA5 spatial resolution in complex terrain, and the need for probabilistic modelling are explicitly acknowledged.

Large-scale governance of urban and periurban green spaces demands perception data that is timely, multi-site, and cost-effective — requirements that established social science methods are structurally ill-equipped to meet. This paper introduces Environmental Digital Perception (EDP), a concept that treats the spontaneous digital discourse of citizens, media, and scientists about natural spaces as a measurable social indicator, and presents an integrated analytical framework to operationalise it. The framework couples fine-tuned Transformer models (RoBERTa-BETO for Spanish, SciBERT for English) with Zero-Shot Natural Language Inference for aspect-level ecosystem service attribution, a bilingual domain lexicon grounded in CICES v5.1, and Inverse Distance Weighting geospatial interpolation. Deployed on 264,481 records from five digital source types spanning Cundinamarca, Colombia (2020–2025), the pipeline achieves an overall accuracy of 86.4% (macro F1 = 0.850, Cohen's κ = 0.72). A 13-dimension structured comparison positions EDP as a complement — not a substitute — to conventional survey methods, delineating the governance contexts in which each is optimal. Results document: (i) a structurally dominant neutral discourse (~75%) reinterpreted as a mobilisation gap rather than social indifference; (ii) a perception–function decoupling in which experiential ecosystem services capture public affect while ecologically critical regulatory services remain digitally invisible; (iii) institutional governance quality as the foremost predictor of negative sentiment, outranking contamination; and (iv) a statistically significant five-year discourse shift from recreational to climate-regulatory framing (Mann-Kendall τ = +0.91, p = 0.001). Geographic disaggregation into three territorial zones reveals distinct EDP profiles linked to biome type, urbanisation pressure, and governance quality. A three-tier institutional integration model renders EDP deployable within existing environmental governance architectures at daily, quarterly, and annual operational rhythms.

In recent years, there has been growing concern about deforestation in the Amazon River basin, particularly in relation to its impact on regional water resources. This study performs a spatial and temporal analysis of deforestation variations between 2001 and 2020, using MODIS and Sentinel data. Using supervised classification techniques, we classified land changes into afforestation, deforestation and reforestation and analyzed the transitions between different land uses including forest, pasture, shrubland, crops and urbanization. Our findings show an annual forest loss of about 5,726 square kilometers, much of this deforestation is localized to specific subregions, although the overall size of the watershed suggests a lack of sensitivity.These results show a significant correlation with variations in evapotranspiration, estimated through a model calibrated with data from the ERA5 reanalysis. This dataset was used to analyze the standardized precipitation and evapotranspiration indexes (SPEI), revealing that, during the last 20 years under study, the region experienced an increase in both the magnitude and intensity of drought compared to the previous 20 years.In particular, a direct relationship is observed between aggressive agricultural policies in Brazil and Bolivia and increased deforestation rates. In addition, this study serves as the basis for complementary research work assessing the implications of these land use changes on river discharge and estimated groundwater recharge in the Amazon basin.

The atmospheric dynamics of the Amazon, critical for global environmental stability, have faced increasing influence from numerous El Niño and La Niña events in recent decades. While reanalysis data has incorporated these events through models and measurements, the intricate mechanics of spatial and temporal water vapor transport remain unclear. In this study, we present a preliminary analysis of these dynamics, utilizing over twenty years of ERA5 monthly data over atmospheric layer. Our investigation was constructed on two primary scales, each offering unique insights. The first scale aims to replicate and validate the system's seasonality concerning the Intertropical Convergence Zone (ITCZ), those patterns allow evaluating some patterns and its effects in land hydrological process observed along the basin integrating specific methods and models to clarify how the seasonality was replicated and validated. On the second scale, we delve into smaller hydrological sub-units of the Amazon, identifying their contribution to water recycling, and net fluxes across the basin boundaries. We provide an innovative estimation of transport paths using a 4 cardinal directional approach (brubaker box scheme modified) that makes possible identificate, analyse and understanding, water sources and sinks and its relevance in the normal hydrological production and synergically systems The findings indicate the system's relatively stable dynamics in terms of water vapor sources and altitudinal variation across atmospheric layers. Our methodology introduces a novel framework for calculating comprehensive trajectories of water vapor transport from a hybrid lagrangian-eulerian approach, significantly enhancing our understanding of the Amazon's hydrological cycle from an atmospheric perspective. To provide more precision, we specify that the stability observed in the system pertains to water vapor sources and altitudinal variation. These stable dynamics contribute valuable insights into the intricate water vapor transport mechanisms in the Amazon. Additionally, we highlight the implications of our findings for future research in understanding how to create alternatives to mitigate some impacts of El Niño and La Niña events on the Amazon's atmospheric dynamics.