ForestSignal-CA: a 30 m, 41-year Landsat-derived data cube of California forest, shrubland, and grassland properties. 19 layers covering vegetation, hydrology, fire hazard, carbon, dieoff risk, and disturbance severity. Created using a unified methodology that emphasizes temporal and cross-property consistency, to track change, explore tradeoffs, and support decision-making and scientific discovery. CC BY-NC 4.0.
This repository contains the first full upload of the ForestSignal-CA data cube — 19 layers × 41 years (water years 1985–2025). Documentation, STAC catalog, integration demos, and the broader release infrastructure are still being finalized over the coming weeks. A CONUS release will follow. Layout and metadata conventions may change before the formal release; contact mgoulden@uci.edu.
ForestSignal develops 40-year+ data cubes of forest, shrubland, and grassland properties, including vegetation structure, fuel loads, fire hazard, biomass, hydrology, and disturbance history. The data are based on Landsat and use a unified approach that emphasizes temporal and cross-property consistency, with a goal of tracking change and exploring tradeoffs. The data are built and openly shared by university scientists to support decision-making and scientific discovery.
19 layers covering water years 1985–2025 (October–September), 30 m resolution, EPSG:5070 (NAD83 / Conus Albers). All files are Cloud-Optimized GeoTIFFs (COGs).
Vegetation cover and structure
Veg_TreeFrac/ — fractional tree canopy coverVeg_ShrubFrac/ — fractional shrub coverVeg_HerbFrac/ — fractional herbaceous coverVeg_BareFrac/ — fractional bare/non-vegetated coverVeg_CanopyHt/ — tree canopy heightHydrology
WaterFlux_AETmax/ — potential evapotranspirationWaterFlux_AETrealized/ — actual evapotranspirationWaterFlux_Soilmoisture/ — soil moistureWaterFlux_Runoff/ — runoffWaterFlux_PminusETmax_SPI0/ — climatological water deficit at SPI-48 = 0Fire hazard
Fire_LCP/ — 8-band landscape file (FARSITE v.4 format) for fire-behavior modelingFire_FlamMap_FL/ — FlamMap-predicted flame lengthFire_FlamMap_ROS/ — FlamMap-predicted rate of spreadCarbon sequestration
Carbon_AGB/ — above-ground biomassCarbon_GPP/ — gross primary productionForest dieoff risk
Vulner_TreeDieoff_SPI-2/ — drought-driven tree dieoff vulnerability under SPI-48 = −2Disturbance severity (water years 1986–2024; first and last years cannot be computed)
Disturbance_TreeFrac/ — annual tree-canopy lossDisturbance_ShrubFrac/ — annual shrub-canopy lossDisturbance_AGB/ — annual above-ground biomass lossForestSignal_CA_Documentation.pdf in this repo contains complete layer descriptions, unit conventions, methodology summaries, projection details, and use limitations. Read it before working with the data quantitatively — several layers (Fire_LCP bands, Disturbance layers, Veg fractions) use specific scaling conventions that affect any numeric analysis.
All files are Cloud-Optimized GeoTIFFs in EPSG:5070, 30 m. Three ways to use them, in increasing order of sophistication.
The COG format means rasters can be read directly from cloud storage over HTTP, transferring only the bytes needed for the current view. The URI form differs between ArcGIS Pro and open-source GIS tools.
For ArcGIS Pro — use the S3 URI, for example:
s3://us-west-2.opendata.source.coop/forestsignal/forestsignal-ca/Carbon_AGB/ForestSignal_CAwide_Carbon_AGB_2025_V260426.tif
In ArcGIS Pro: Add Data → From Path → paste the URI.
For QGIS, GDAL, rasterio, terra — use the /vsicurl/ URL, for example:
/vsicurl/https://data.source.coop/forestsignal/forestsignal-ca/Carbon_AGB/ForestSignal_CAwide_Carbon_AGB_2025_V260426.tif
In QGIS: Layer → Add Layer → Add Raster Layer → paste the URI as the Source.
Both URI forms point at the same underlying file. Source Cooperative's file detail pages show both forms — "Cloud URI" is the S3 form (for ArcGIS), "Source URL" is the HTTPS form (prepend /vsicurl/ for GDAL-based tools). Substitute the layer name and year for any other file.
Every file has a public URL you can fetch with any HTTP client. No credentials needed.
https://data.source.coop/forestsignal/forestsignal-ca/Carbon_AGB/ForestSignal_CAwide_Carbon_AGB_2025_V260426.tif
Use curl, wget, browser save, Invoke-WebRequest on Windows PowerShell.
For downloading multiple files or whole layer directories, the AWS CLI is fastest. No credentials or AWS account needed — use the --no-sign-request flag:
AWS CLI is available for Windows, Mac, and Linux at aws.amazon.com/cli.
Files follow the convention ForestSignal_CAwide_{LayerName}_{YYYY}_V{YYMMDD}.tif, where {YYYY} is the four-digit water year and V{YYMMDD} is the processing-batch date. The version stamp is provisional and may move to STAC metadata in the final release.
These data are shared under the Creative Commons Attribution-NonCommercial 4.0 International license (CC BY-NC 4.0). Use of these data implies consent to the license terms and to the University of California disclaimer reproduced in §6 of the documentation PDF and summarized below. The license restricts use for commercial purposes — see creativecommons.org/faq/ for a discussion of commercial use.
The data are offered with no promise of technical support and with the hope that users will provide feedback to help refine and improve them.
The materials are offered "as is"; the user assumes any and all risks of any kind or amount in using these materials. Access or use of these data implies consent to the disclaimer in the dataset documentation (§6 of ForestSignal_CA_Documentation.pdf), which is the binding statement.
Goulden, M.L. (2026). ForestSignal CA: 41 years of reconciled forest dynamics for California, water years 1985–2025 (Version V260426). Source Cooperative. [DOI to be assigned]
mgoulden@uci.edu — notes on errors and use cases are especially valuable. Feedback will materially shape the final release.
Preliminary release V260426 (April 2026 build). University of California, Irvine.