From 50d5a2b27f1d2e40eb37ec5910638c5d655a0e51 Mon Sep 17 00:00:00 2001 From: Yangshuang Xu Date: Tue, 6 Jan 2026 00:37:43 -0500 Subject: [PATCH] Update pyhazards_datasets.rst.txt Refine dataset descriptions by adding update frequency, contents, and task relevance. Reformat entries to ensure compatibility with Sphinx list-table rendering. Signed-off-by: Yangshuang Xu --- docs/_sources/pyhazards_datasets.rst.txt | 21 ++++++++++++++------- 1 file changed, 14 insertions(+), 7 deletions(-) diff --git a/docs/_sources/pyhazards_datasets.rst.txt b/docs/_sources/pyhazards_datasets.rst.txt index 36c376b0..544d8536 100644 --- a/docs/_sources/pyhazards_datasets.rst.txt +++ b/docs/_sources/pyhazards_datasets.rst.txt @@ -14,18 +14,25 @@ Datasets :header-rows: 0 :class: dataset-list - * - ``MERRA-2`` - - MERRA-2 is a global weather and climate dataset created by NASA that provides long-term records of atmospheric conditions from 1980 to the present. The data are organized on a regular global grid with hourly to monthly updates and a spatial resolution of about 0.5° × 0.625°, and include variables such as temperature, wind, humidity, surface fluxes, and aerosols, making it useful for large-scale climate and hazard analysis worldwide. + * - ``MERRA-2 (NASA GMAO)`` + - Data format: commonly distributed as NetCDF (and also accessible via NASA services); data structure: gridded arrays (tensor-like) with dynamic surface fields, vertical profile fields, and static layers; dimensions: time is the temporal index (hourly), latitude and longitude are the spatial grid axes (about 0.5 deg x 0.625 deg), and vertical variables include an additional pressure-level axis; update frequency: released with a typical latency of a few weeks after real time; tasks: provide core meteorological drivers for weather/climate forecasting (e.g., WxC-style pipelines) and support hazard prediction studies such as wildfire and hurricane by linking hazard outcomes to atmospheric and land-surface conditions; link: `MERRA-2 + overview `_; reference: `Gelaro et al. (2017) `_. * - ``ERA5`` - - ERA5 is a global reanalysis dataset produced by ECMWF and is commonly used as a reference for weather and climate studies. It covers the period from 1950 to the present and provides hourly data on a global grid at roughly 0.25° resolution, including precipitation, temperature, wind, and land-surface variables, which allows detailed environmental and hazard modeling at a global scale. + - Data format: commonly distributed as GRIB and NetCDF via Copernicus CDS; data structure: gridded arrays (tensor-like) with single-level fields and optional pressure/model-level fields; dimensions: time is the temporal index (hourly), latitude and longitude are the spatial grid axes (commonly 0.25 deg x 0.25 deg on CDS products), and vertical products include an additional level axis; update frequency: updated daily with a short-latency stream (ERA5T) and later consolidated into the final ERA5; tasks: support weather/climate prediction and benchmarking, and serve as standardized meteorological covariates for hazard prediction such as wildfire danger, heavy-rain/flooding, and hurricane-related environmental conditions; link: `ERA5 + single levels `_; reference: `Hersbach et al. (2020) + `_. * - ``NOAA Flood Events (Storm Events Database)`` - - The NOAA Flood Events dataset collects reported flood events across the United States, including when and where floods occurred and the impacts they caused. The data are stored as event-based records rather than grid data, usually at county or local levels, and span many decades, making them suitable for analyzing flood patterns and risks within the U.S. + - Data format: commonly distributed as tabular text (CSV) and database extracts; data structure: event-based records (not tensors) where each row is an event report with fields rather than gridded axes; dimensions/fields: time fields represent event begin/end timestamps, location fields represent reporting geography (often county/zone and coordinates where available), and impact fields represent reported damages and casualties; update frequency: updated on a rolling basis with a typical delay of about 75–90 days after the end of the data month; tasks: support flood occurrence/impact analysis, event frequency and trend studies, and supervised modeling where the target is event occurrence or impact rather than meteorological state; link: `Storm Events + Database `_; reference: `NCEI metadata record `_. * - ``FIRMS (Fire Information for Resource Management System)`` - - FIRMS (Fire Information for Resource Management System) provides near-real-time satellite observations of active fires around the world. The dataset consists of time-stamped fire locations detected by MODIS and VIIRS sensors at spatial resolutions between about 375 m and 1 km, and is widely used for global wildfire monitoring and analysis. + - Data format: commonly distributed as tabular and GIS-ready formats (CSV, Shapefile/GeoJSON/KML) depending on access method; data structure: event/point detections (not tensors) where each record is a time-stamped fire/hotspot detection; dimensions/fields: time is detection timestamp, location is latitude/longitude (point geometry), and attributes include sensor/product fields such as confidence and radiative measures (product-dependent); update frequency: near-real-time publication refreshed multiple times per day, with detections typically available within hours of satellite overpass; tasks: operational wildfire monitoring/nowcasting and rapid situational awareness, and serving as event labels/targets for wildfire prediction pipelines when combined with meteorology and fuels/land layers; link: `FIRMS + portal `_; reference: `VIIRS active fire product paper `_. * - ``MTBS (Monitoring Trends in Burn Severity)`` - - MTBS (Monitoring Trends in Burn Severity) is a dataset that maps wildfire burn areas and severity across the United States. It focuses on large fires since 1984 and uses Landsat satellite imagery at 30 m resolution, making it useful for studying long-term wildfire impacts and regional fire risk. + - Data format: commonly distributed as GIS rasters and vectors (e.g., GeoTIFF and shapefiles/geodatabases); data structure: mapped fire perimeters and burn severity layers (not a single tensor) where products are organized by fire event and can be aggregated by year; dimensions/fields: spatial axes represent 30 m raster grids for severity layers and polygon geometries for perimeters, and time is represented by fire year/event metadata; update frequency: produced on an annual cycle as new fire-year products are completed; tasks: post-fire impact assessment, long-term wildfire regime/trend analysis, and providing training/validation layers for models that predict burn severity/extent or relate severity to meteorology, fuels, and topography; link: `MTBS portal `_; reference: `Eidenshink et al. (2007) + `_. * - ``LANDFIRE Fuel (USFS / LANDFIRE)`` - - LANDFIRE provides fuel and vegetation information across the United States to support wildfire modeling and risk assessment. The data are distributed as raster layers at about 30 m resolution and describe fuel models, canopy structure, and vegetation types, helping researchers and practitioners understand and simulate wildfire behavior at landscape scales. + - Data format: commonly distributed as GIS rasters (e.g., GeoTIFF) and supporting metadata; data structure: gridded map layers (raster matrices) describing vegetation and fuels at national scale; dimensions/fields: spatial axes represent ~30 m raster grids, time is represented by product version/year (and disturbance layers where applicable), and values encode vegetation/fuel attributes such as type, cover, height, and fuel descriptors; update frequency: released as versioned updates and commonly treated as an annual updating program in applied workflows; tasks: fuels/vegetation characterization for wildfire behavior simulation, landscape-scale wildfire risk assessment, and static covariates for wildfire occurrence/spread models paired with dynamic meteorological predictors; link: `LANDFIRE data access + `_; reference: `LANDFIRE program overview `_. Core classes ------------