Climate Toolkit for Agroecological Research
1 Introduction
The Alliance of Bioversity International and CIAT (ABC), in partnership with the McKnight Foundation and the African Institute of Mathematical Sciences (AIMS), is developing a Climate Data Toolkit to support agroecological (AE) research. The toolkit will empower researchers and practitioners to integrate climate data into their workflows, enabling robust analyses of agroecological systems under climate stresses and shocks. Rather than creating new datasets or analytical functions, this project will consolidate and simplify access to existing climate data resources, offering users a streamlined, user-friendly platform to support evidence-based decision-making.
This document is organized into several key sections to guide users through the Climate Data Toolkit. It begins with an overview of the climate datasets included, detailing their sources, variables, and formats. Next, it presents the climate indicators relevant to agroecological research, explaining their definitions and calculation methods. The following section outlines the prioritization methodology used to select and rank datasets and indicators based on research needs. Additional chapters provide information on the tools, scripts, and APIs integrated within the toolkit, along with practical guidance on applying the toolkit in real-world agroecological contexts. Finally, references and supplementary materials are included to support further exploration and use.
2 Climate datasets
This section systematically presents the climate datasets used in this study, covering their key characteristics, accessibility procedures, and their specific value for advancing agroecological research. The datasets were evaluated and prioritized based on criteria critical for agricultural applications, data integration, and scientific rigor (Prioritization method).
| Dataset Name | Short Description | Provider/Source | Spatial Coverage | Temporal Coverage | Spatial Resolution | Temporal Resolution | Key Variables | Data Quality Indicators | Metadata Completeness | Update Frequency | Data Provenance | Preprocessing Steps | Access Methods | API/Package Links | Example Scripts | License/Terms | Integration Challenges | Data Ownership | Versioning/Change Log | Data Citation | Benchmarking/Validation Studies | Potential Applications (Agroecological) | Priority Level |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| CHIRPS | High-resolution, quasi-global rainfall dataset blending satellite imagery and station data for trend analysis and seasonal drought monitoring | Climate Hazards Center (CHC), UCSB | 50°S–50°N, 180°W–180°E | 1981–present (updated frequently) | 0.05° x 0.05° (~5.5 km at equator) | Daily, pentadal (5-day), dekadal (10-day), monthly | Precipitation (mm) | Validated against station data; uncertainty estimates; provisional data for ~20 days after month end | Spatial/temporal coverage, source, version, units, coordinate system (WGS84) | Updated within 45 days; provisional data replaced with final | Satellite + station blend; processing steps documented on CHC website | Resampling, masking, unit conversion, missing data handling | FTP/HTTP (CHC data portal), FEWS NET Data Portal, Google Earth Engine (GEE), R package (chirps on CRAN) |
GEE API Docs, R package chirps | Example: Script to Download CHIRPS in R (GitHub) | Open for research and non-commercial use | Large file sizes, need for specific libraries (e.g., netCDF4, xarray), handling provisional data, managing missing data flags |
Climate Hazards Center, UCSB | Version history on CHC website | Cite as: “CHIRPS dataset provided by Climate Hazards Group, UCSB” | CHIRPS Validation on CHC website, CHIRPS v2.0: Increased resolution from climate station evaluation (Funk et al., 2015) | Rainfall pattern analysis, drought monitoring, crop yield modeling, water resource management, agroclimatic zoning, validation of local climate data | High (core dataset for agroecology) |
| AgERA5 | Reanalysis dataset tailored for agriculture, based on ERA5 but with additional variables and adjustments for agricultural applications | ECMWF/C3S | Global | 1979–present (updated daily with a short delay) | 0.1° x 0.1° (~10 km) | Hourly | Standard ERA5 variables plus: Evapotranspiration (actual/potential), Soil water content (various depths), Leaf Area Index (LAI), Fraction of Absorbed Photosynthetically Active Radiation (FAPAR), crop-specific indicators | Model-based, validated against ground observations where possible; uncertainty and bias information in documentation; see ECMWF documentation | Full metadata in GRIB files (variable definitions, units, provenance, coordinate system, version) | Updated daily with a short delay | Derived from ERA5 reanalysis; further processed for agricultural relevance; see AgERA5 documentation | Convert GRIB to NetCDF for easier handling (cfgrib, eccodes, xarray); spatial/temporal subsetting; handle large file sizes; check units/definitions for ag variables |
Copernicus Data Space Ecosystem (CDSE), ECMWF AgERA5 page | CDS API Docs, cfgrib Python package, eccodes | Example: Script to download AgERA5 in R(GitHub) | Copernicus license (free for research/educational use, registration required) | Registration required, GRIB format requires specific tools, large data volume, agricultural variable definitions may differ from other datasets | ECMWF, Copernicus Climate Change Service | Versioning and change log available on CDSE and ECMWF AgERA5 | Cite as: “AgERA5 dataset provided by ECMWF and Copernicus Climate Change Service (C3S)” | AgERA5: Dataset evaluation and use for crop modeling (ECMWF), AgERA5 technical note (PDF) | Crop yield forecasting, irrigation management, pest/disease modeling, land suitability assessment, drought/heat stress monitoring, vegetation dynamics (LAI, FAPAR), water resource management, climate change impact assessment, soil moisture studies | High (core dataset for agroecology and climate-adaptive agriculture) |
| TerraClimate | High-resolution (~4 km, 1/24°) monthly global dataset of climate and climatic water balance variables for terrestrial surfaces, combining climatological normals with coarser time-varying data for detailed, temporally consistent records | Climatology Lab, University of California Merced (UCM) | Global (all terrestrial surfaces) | 1958–present (updated annually) | 1/24° (~4 km) | Monthly | Precipitation, max/min temperature, wind speed, vapor pressure, vapor pressure deficit, downward shortwave radiation, reference/actual evapotranspiration, climatic water deficit, soil moisture, runoff, snow water equivalent, Palmer Drought Severity Index (PDSI) | Validated against station and streamflow data; improved mean absolute error and spatial realism; inherits biases from parent datasets (e.g., precipitation bias in mountains) | Metadata available for variables, units, provenance, and grid; NetCDF metadata includes time, lat, lon, variable | Updated annually when parent datasets become available | WorldClim v2 (climatology), CRU Ts4.0 and JRA55 (monthly anomalies), climatically aided interpolation; water balance via modified Thornthwaite-Mather model | Check missing data flags; handle large files; use geospatial tools (netCDF4, xarray, raster, R) | Climatology Lab website, Microsoft Planetary Computer, Google Earth Engine, THREDDS servers | Google Earth Engine API Docs, Planetary Computer API Docs, THREDDS server access, R package (datazoom.amazonia vignette) | Download TerraClimate in Python/R via THREDDS server, or use GEE script example | Public domain (CC0 license); free for research and operational use | Large file sizes; requires geospatial software; annual updates; monthly resolution limits capture of short-term extremes; model-derived variables have uncertainties in data-sparse regions | Climatology Lab, University of California Merced | Versioning/change log not specified; annual updates tracked via parent datasets | Abatzoglou et al. (2018): TerraClimate, a high-resolution global dataset of monthly climate and climatic water balance from 1958–2015. Scientific Data, 5, https://doi.org/10.1038/sdata.2017.191 | Scientific Data article (Abatzoglou et al., 2018), Climate Data Guide summary | Climate trend/variability analysis, drought monitoring, water balance and agricultural risk assessment, crop yield modeling, agroclimatic zoning, ecological/hydrological modeling, validation/downscaling of coarser datasets | High (core dataset for agroecological and climate-adaptive research) |
| IMERG (Integrated Multi-satellitE Retrievals for GPM) | NASA/JAXA quasi-global (90°S–90°N), high-resolution (0.1° x 0.1°) precipitation estimates by merging GPM satellite constellation data with microwave, infrared, and gauge observations | NASA GPM / JAXA | Quasi-global (90°S–90°N, 180°W–180°E) | June 2000–present (updated with latency: Early, Late, Final Runs) | 0.1° x 0.1° (~10 km) | Half-hourly, daily, monthly | Precipitation rate (mm/hr), accumulated precipitation (mm), error estimates, counts of contributing retrievals | Validation against global gauge networks; error estimates included; Final Run is highest quality; see IMERG Algorithm Theoretical Basis Document | NetCDF files include variable/unit metadata, time, lat, lon, run type, version, and error variables | Half-hourly to monthly, depending on product and run (Early: ~4h latency; Late: ~12h; Final: ~3 months) | Merges GPM satellite data (microwave/IR sensors) with gauge data; multiple runs differ by latency and data sources; see IMERG documentation | Handle large files; use netCDF4/xarray/rioxarray (Python) or raster (R); check fill values; aggregate temporally as needed; select appropriate run (Early/Late/Final) | NASA GES DISC, Google Earth Engine, ERDDAP servers (example) | GEE API Docs, GES DISC Data Access, IMERG Data Cookbook | GES DISC IMERG Python Example, GEE JavaScript Example | NASA Data Policy (open for research/non-commercial use, acknowledge NASA/JAXA/GPM) | Large file sizes; software requirements; different runs have different latency/accuracy; missing data flags; only precipitation (no temperature, etc.) | NASA / JAXA | IMERG Versioning (version and change log per product) | “Huffman, G.J., et al., 2020. NASA Global Precipitation Measurement (GPM) Integrated Multi-satellitE Retrievals for GPM (IMERG), Version 07. NASA’s GES DISC. https://doi.org/10.5067/GPM/IMERG/3B-HH/07” | IMERG Validation Reports, IMERG Algorithm Theoretical Basis | Rainfall pattern analysis, drought monitoring, extreme precipitation event analysis, crop yield modeling, water resource management, agroclimatic zoning, validation of local rainfall data, sub-daily/daily hydrological modeling | High (core precipitation dataset for agroecological and climate studies) |
| TAMSAT | Long-term, high-resolution rainfall dataset for Africa, blending Meteosat thermal infrared satellite imagery with ground-based rain gauge observations for operational rainfall monitoring and climate risk assessment. | Department of Meteorology, University of Reading | Africa (continental) | 1983–present (updated regularly) | 0.0375° x 0.0375° (~4 km) | Daily, dekadal (10-day), monthly | Rainfall/precipitation (mm), soil moisture estimates (for some products) | Calibrated with ground-based rain gauges; validation and uncertainty assessments published in peer-reviewed literature; quality control flags in data files | NetCDF files include variable names, units, time, lat, lon, version, and quality flags; documentation available on TAMSAT website | Updated regularly as new satellite and gauge data become available | Derived from Meteosat TIR imagery, calibrated with gauge observations; see TAMSAT documentation and publications | Handle large files; use NetCDF-compatible tools (e.g., xarray, netCDF4, raster); check for missing data flags and quality control layers; aggregate or subset as needed | TAMSAT Data Portal, FTP/HTTP download (see website for latest), may require registration | No formal public API; use TAMSAT Data Portal for downloads; process with xarray, netCDF4 (Python), or raster (R); check documentation for updates |
- | Terms of Use: Free for non-commercial research and educational use; citation required | Large file sizes; NetCDF format requires specific tools; access methods may change; rainfall only (no temperature, radiation, etc.); accuracy limited in regions with sparse gauge data | University of Reading | Versioning and updates noted in file names and documentation | “TAMSAT rainfall dataset, University of Reading. See TAMSAT website for citation guidance.” | TAMSAT Publications, validation studies listed on the website | Rainfall pattern analysis, drought monitoring and early warning, agricultural planning, crop yield modeling, water availability assessment, climate risk analysis, validation of other rainfall datasets over Africa | High (core rainfall dataset for agroecological research in Africa) |
| CHIRTS (Climate Hazards Center InfraRed Temperature with Stations) | High-resolution, quasi-global maximum and minimum temperature dataset blending satellite thermal infrared data with ground station observations for climate monitoring and agricultural risk assessment | Climate Hazards Center (CHC), UCSB | 60°S–70°N, 180°W–180°E | 1983–2016 (v1.0), 1983–present (v2.0, ongoing updates) | 0.05° x 0.05° (~5.5 km at equator) | Daily | Maximum and minimum temperature (°C) | Validated against global station data; quality flags and uncertainty estimates included; see CHIRTS documentation | Metadata includes variable, unit, time, lat, lon, version, and processing details | Updated as new versions or extensions are released (check CHC website for latest) | Blends satellite thermal infrared observations with ground station data; processing described in CHIRTS documentation | Handle NetCDF files; use xarray, netCDF4 (Python), or raster (R); check for missing data flags | CHC Data Portal (CHIRTS), FTP/HTTP download (see website) | No formal API; use CHC data portal for downloads and documentation | Example: Download and process CHIRTS in R (GitHub) | Open for research and non-commercial use (see CHC data use policy) | Large file sizes; requires geospatial software; limited to temperature variables | Climate Hazards Center, UCSB | Version history and change log on CHC website | “CHIRTS dataset provided by Climate Hazards Group, UCSB” (see CHC citation guidance) | Validation and methodology described in CHIRTS documentation and associated peer-reviewed publications | Heat stress monitoring, crop modeling, climate trend analysis, agricultural risk assessment, drought impact studies, validation of other temperature datasets | High (core temperature dataset for agroecological research) |
| ERA5 (ECMWF Reanalysis v5) | Fifth generation ECMWF atmospheric reanalysis, providing hourly estimates for a large number of atmospheric, land, and ocean climate variables globally | European Centre for Medium-Range Weather Forecasts (ECMWF) | Global | 1950 to near-present (updated daily) | 0.25° x 0.25° (~31 km at equator) for atmospheric variables; other variables may differ | Hourly (monthly means also available) | Air temperature (various levels), precipitation, surface radiation, wind speed/direction, soil moisture, evaporation, sea surface temperature, sea ice, and more | Model-based with uncertainties estimated via ensemble; validated against observations; see ERA5 documentation | Full metadata in GRIB files (variable definitions, units, provenance, coordinate system, version) | Updated daily with a delay of 2–3 months for final data | Combines model data with global observations using 4D-Var data assimilation; see ERA5 documentation[1][3][5] | Use cfgrib/eccodes (Python) for GRIB; convert to NetCDF; subset spatially/temporally; check units/definitions | Copernicus Data Space Ecosystem (CDSE), ECMWF Data Catalogue (MARS), Cloud Platforms (AWS, GCP, Azure) | CDS API Docs, cfgrib Python package, eccodes | ERA5 Python example (CDS API) | Copernicus license (free for research/educational use, registration required) | Registration required; large dataset volume; GRIB format needs specific tools; high computational demands; complexity of access and format | ECMWF, Copernicus Climate Change Service (C3S) | Versioning and change log available on ECMWF documentation | “ERA5 data provided by ECMWF and Copernicus Climate Change Service (C3S)” | ERA5 evaluation and publications (ECMWF), Hersbach et al. (2020) | Temperature analysis, precipitation/drought monitoring, evapotranspiration, solar radiation, wind patterns, soil moisture, extreme weather, climate change impact studies | High (core dataset for agroecological and climate research) |
| NASA Earth Exchange Global Daily Downscaled Projections (NEX-GDDP) | Global, high-resolution (0.25°) daily climate projections downscaled from CMIP5 and CMIP6 GCMs, including key variables for climate impact studies | NASA Earth Exchange (NEX) | CMIP5: 50°S–50°N; CMIP6: 60°S–90°N | CMIP5: 1950–2099; CMIP6: 1950–2100 | 0.25° x 0.25° (~25 km) | Daily | tasmin (daily min temperature, K), tasmax (daily max temperature, K), precipitation (kg m⁻² s⁻¹) | Bias-corrected and spatially disaggregated; multiple GCMs/scenarios; see NEX-GDDP Technical Notes (CMIP5), NEX-GDDP-CMIP6 Technical Note | NetCDF metadata includes variable, unit, scenario, model, time, lat, lon, and processing info | Dataset is not updated in real time; new versions may be released for CMIP6/other projects | Downscaled from CMIP5/CMIP6 GCMs using Bias-Correction Spatial Disaggregation (BCSD); see technical documentation (CMIP5), NEX-GDDP-CMIP6 Technical Note | Handle large NetCDF files; use xarray, netCDF4, CDO; check for missing data flags; convert units as needed | NASA NCCS THREDDS, AWS S3, Google Earth Engine | Google Earth Engine API Docs, THREDDS NetCDFSubset, AWS S3 API | GEE example scripts, AWS S3 download example, GitHub script to download data | Open for scientific research; not recommended for commercial/engineering use without expert consultation | Large file sizes (up to 38 TB for CMIP6), requires technical expertise, limited to temperature/precipitation, projections subject to GCM/downscaling uncertainties | NASA Earth Exchange (NEX) | Versioning and change log available via NEX-GDDP Technical Notes (CMIP5), NEX-GDDP-CMIP6 Technical Note | “NASA Earth Exchange Global Daily Downscaled Projections (NEX-GDDP)” (see homepage for citation details) | NEX-GDDP Technical Notes (CMIP5), NEX-GDDP-CMIP6 Technical Note | Climate risk assessment, crop yield modeling, drought/heat stress analysis, water resource planning, agroclimatic zoning, validation of local climate data | High (core dataset for climate impact and agroecological research) |
| NASA POWER (Prediction Of Worldwide Energy Resources) | Freely available, gridded solar and meteorological datasets derived from NASA satellite observations and climate models for renewable energy, agriculture, and building energy efficiency applications | NASA Langley Research Center (LaRC) POWER Project | Global | Varies by parameter; many from 1981 to near-present | 0.5° x 0.5° (~50 km); some parameters may be higher resolution | Daily, monthly, hourly (parameter-dependent) | Solar radiation (insolation, direct normal irradiance, diffuse horizontal irradiance), surface meteorology (temperature, wind speed/direction, humidity, pressure), precipitation, evapotranspiration, soil moisture (some datasets), cloud cover | Derived from NASA satellite products and reanalysis; validation and uncertainty information in documentation; see POWER Documentation | Metadata included in NetCDF, JSON, and CSV outputs; variable names, units, coordinates, time, and processing info | Updated regularly as new satellite and reanalysis data are processed | Combines satellite observations and climate models; see POWER Documentation for details | Handle missing data flags; use appropriate tools for CSV, JSON, NetCDF (e.g., xarray, pandas, netCDF4); check units and spatial/temporal resolution | POWER Data Access Viewer, POWER API, direct download (CSV, JSON, NetCDF, GeoJSON) | API Getting Started, POWER Python client (pynasapower), POWER R client (nasapower) | API Python Example, R Example, API Tutorials | Public domain; free for use with attribution to NASA/POWER (citation guidance) | 0.5° resolution may be too coarse for local studies; temporal coverage varies by parameter; API requires programming knowledge; handle units and missing data carefully | NASA Langley Research Center (LaRC) | Versioning and updates documented in POWER Documentation | NASA POWER Citation Guidance | POWER Publications, Validation/Uncertainty | Solar resource assessment, crop modeling, irrigation planning, climate risk analysis, energy efficiency studies, agroclimatic zoning, weather-driven agricultural decision support | High (widely used for agroclimatology and renewable energy) |
| CMIP6 (Coupled Model Intercomparison Project Phase 6) | International coordinated climate modeling effort providing a comprehensive ensemble of global climate projections for understanding past, present, and future climate change | World Climate Research Programme (WCRP) WGCM, data via Earth System Grid Federation (ESGF) | Global (resolution varies by model) | Historical: ~1850–present; Projections: up to 2100+ (varies by experiment) | Varies by model (from ~100 km to ~25 km or finer) | Daily, monthly, sub-daily (model/variable-dependent) | Temperature (surface, ocean, etc.), precipitation, radiation, wind, humidity, sea level pressure, sea ice, ocean currents, salinity, soil moisture, runoff, carbon cycle, and more | Each modeling group provides validation and uncertainty info; ensemble approach allows assessment of robustness; see Eyring et al., 2016 | NetCDF files with CF-compliant metadata (variable/unit, model, experiment, grid, version, etc.) | Not updated in real time; new experiments released as available | Data produced by international modeling centers; standardized protocols; see CMIP6 Overview | Handle large NetCDF files; use xarray, netCDF4, dask; understand DRS structure; regrid as needed for analysis | ESGF Search, Copernicus CDS, CEDA STAC API | ESGF API Docs, pyesgf Python library, Copernicus CDS API | Example: Script to download rowCMIP6 in R(GitHub) | Data access and use governed by contributing modeling group terms (see ESGF Terms), generally free for research/education with attribution | Huge data volume; variable model resolutions; complex DRS structure; requires technical expertise; model biases and uncertainties; license terms vary by group | WCRP WGCM, Modeling Centers | Versioning and change log per model/experiment; see CMIP6 DRS | See CMIP6 Citation Guidance | Eyring et al., 2016, CMIP6 Publications | Climate change impact assessment, crop suitability, extreme event analysis, adaptation strategy development, regional downscaling, climate risk analysis | High (core dataset for climate and agroecological research) |
3 Indicators
This section summarizes key climate hazards: drought, heat stress, waterlogging, and compound events, relevant to agroecological research and climate risk assessment. For each hazard, widely used indices and indicators are listed along with their calculation methods, interpretation thresholds (including crop or livestock specificity where applicable), data requirements, and available tools or scripts for computation. Compound hazards, such as simultaneous heat and drought, are also included, as they can amplify risks to crops and livestock. This resource is designed to help users select, calculate, and apply the most appropriate indices for climate risk analysis and decision support in agricultural systems.
| Hazard | Type/Subtype | Indicator/Index/Variable | Calculation/ Description | Thresholds / Interpretation | Crop/Livestock Specificity | Data Needed / Source | Script/API/Package/Link | Reference / Link | Soil Data Needed | Elevation Data Needed | Prioritization | Example Application |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Drought | Meteorological | SPI (Standardized Precipitation Index) | SPI = (precip - mean)/std (over ref period) | SPI < -1 (moderate), < -2 (severe) | Generic | Precipitation (monthly; CHIRPS, TAMSAT, IMERG) | R: SPEI package | McKee et al., 1993 | - | - | High | Drought monitoring, early warning |
| Drought | Agricultural | SPEI (Standardized Precipitation Evapotranspiration Index) | Like SPI but includes PET | SPEI < -1 (mod), < -2 (severe) | Generic | Precipitation, PET (CHIRPS, ERA5, AgERA5) | R: SPEI package | Vicente-Serrano et al., 2010 | - | - | High | Drought impact on crops |
| Drought | Agricultural | Soil Moisture Anomaly | SMA = (soil moisture - mean)/std | Crop-specific (e.g. <20% FC) | Yes | Soil moisture (ERA5, AgERA5, TAMSAT) | Python: xarray | FAO | Soil texture, field capacity, depth | - | High | Crop stress detection |
| Drought | Hydrological | Streamflow Anomaly | Deviation from mean streamflow | < threshold (site-specific) | Yes | River gauge data | Custom (site-specific) | WMO | - | Elevation (slope, drainage patterns) | Medium | Water resource management |
| Heat Stress | Crop | NTX (Number of Days Tmax > X°C) | Count days Tmax > threshold during season | X = crop-specific (e.g. 35°C maize, 30°C potato) | Yes | Daily Tmax (CHIRTS, ERA5, AgERA5) | Python Example | Lobell et al., 2011 | - | - | High | Heatwave risk for maize |
| Heat Stress | Livestock | THI (Temperature Humidity Index) | THI = T - [(0.55 - 0.0055RH)(T-14.5)] | THI > 72 (dairy cattle stress) | Yes | Temp, humidity (ERA5, AgERA5) | R: climwin | NOAA | - | - | Medium | Dairy cattle management |
| Heat Stress | Crop/Livestock | VPD (Vapor Pressure Deficit) | VPD = SVP - AVP (kPa) | Crop/livestock-specific | Yes | Temp, humidity (ERA5, AgERA5, CHIRTS) | Python: metpy | FAO | - | Elevation (altitude affects temperature) | Medium | Crop/livestock risk mapping |
| Heat Stress | Crop | Hot Nights (NTN) | Number of nights Tmin > threshold (e.g. 20°C) | Crop-specific (e.g. rice, >24°C) | Yes | Daily Tmin (CHIRTS, ERA5, AgERA5) | Python: pandas | Peng et al., 2004 | - | - | Medium | Rice yield loss risk |
| Heat Stress | Crop | Growing Degree Days (GDD) | GDD = Σ(Tmean - Tbase) over season | Tbase = crop-specific | Yes | Daily Tmin, Tmax (CHIRTS, ERA5, AgERA5) | R: chillR | FAO | Soil type (affects crop phenology) | - | High | Crop phenology modeling |
| Waterlogging | - | Days Soil Moisture > Field Capacity | Count days soil moisture > FC | >3 days = risk for many crops | Yes | Soil moisture (ERA5, AgERA5, TAMSAT) | Python: xarray | FAO | Soil porosity, bulk density | - | Medium | Rice, wheat management |
| Waterlogging | - | Days Rainfall > X mm | Count days rainfall > threshold | X = crop-specific | Yes | Precipitation (CHIRPS, TAMSAT, IMERG) | Python: pandas | FAO | - | - | Medium | Flood risk assessment |
| Compound | Crop/Livestock | Heat + Drought (e.g. VPD) | VPD = SVP - AVP (kPa), combines heat and dryness | Crop/livestock-specific | Yes | Temp, humidity (ERA5, AgERA5, CHIRTS) | Python: metpy | FAO | - | - | Medium | Compound hazard assessment |
Key Datasets, APIs, and Software Tools for Soil and Elevation Data
This table provides an overview of the main soil and elevation datasets relevant in our project, along with their spatial resolution, coverage, available APIs, and software packages to facilitate data access and analysis.
| Data Type | Source / Data Portal | Resolution | Coverage | Access Link | Variables Included | API / REST Endpoint | Software Package(s) / Libraries |
|---|---|---|---|---|---|---|---|
| Soil Data | SoilGrids (ISRIC) | 250m | Global | SoilGrids | Soil texture, bulk density, organic carbon, pH, sand, silt, clay fractions | SoilGrids REST API | Python: soilgrids, R: soilDB (partial) |
| Soil Data | OpenLandMap Soil Organic Carbon | 250m | Global | OpenLandMap | Soil organic carbon content | No dedicated public API documented | No widely used package known |
| Soil Data | HWSD (Harmonized World Soil Database) | 1km | Global | FAO HWSD | Soil texture, bulk density, water storage capacity, organic carbon | No official API; data downloadable as GeoTIFF/ASCII | Various GIS software (QGIS, ArcGIS) for data use |
| Elevation | SRTM (NASA) | 30m | Global | Earthdata | Digital elevation model (DEM), slope, aspect | APIs via NASA Earthdata services | Python: elevation |
| Elevation | ALOS World 3D | 30m | Global | ALOS | Digital elevation model (DEM), terrain height | No public API; data downloadable | GIS software, custom scripts |
| Elevation | OpenTopography API | Variable | Global | OpenTopography | Digital elevation data, slope, terrain metrics | OpenTopography REST API | Various GIS and Python libraries |
4 Universal Temperature Thresholds Across Life: Humans, Livestock, Crops, and Fisheries
Article 1: The upper temperature thresholds of life
A growing body of research indicates that the upper temperature thresholds for heat stress and lethality are surprisingly similar across humans, livestock, poultry, major crops, and even many fish species. These thresholds define the comfort or optimal zones for physiological functioning, as well as the points at which moderate, strong, and lethal stress occur.
The table below summarizes the comfort zones and critical temperature thresholds for heat stress and lethality in key organisms, based on a comprehensive review by Asseng et al. (2021). These insights are crucial for understanding the broad impacts of rising temperatures under climate change, as well as for designing effective adaptation strategies in agriculture, food systems, and public health.
| Organism/Group | Comfort/Optimal Zone (°C) | Moderate Stress (°C) | Strong Stress (°C) | Very Strong Stress (°C) | Extreme/Lethal (°C) | Notes / Crop-Specific Details | References |
|---|---|---|---|---|---|---|---|
| Humans | 17–24 | 23–27 | 32 | 36 | >40 (medium humidity) | Lethal: >32 (high humidity), >50 (low humidity), or wet bulb >35°C; vulnerable: children, elderly, ill | Asseng S, Spänkuch D, Hernandez-Ochoa IM, Laporta J. The upper temperature thresholds of life. The Lancet Planetary Health. 2021;5(6):e378-e385. |
| Livestock (Cattle) | 17–24 | 23–27 | 32 | 36 | >40 (medium humidity) | Milk yield drops 10–20% under heat strain; lethal: >35 (high humidity), >40 (low humidity) | |
| Livestock (Pigs) | 17–24 | 23–27 | 32 | 36 | >40 (medium humidity) | Similar to cattle; pigs less efficient at sweating | |
| Poultry (Chickens) | 17–24 | 23–27 | 32 | 36 | >40 (medium humidity) | Reduced egg production, growth, reproduction; lethal: >35 (high humidity), >40 (low humidity) | |
| Maize | 17–24 | 25–30 | >32 | >35 | >40 | Flowering & grain filling most sensitive; yield loss above 35°C at flowering | |
| Wheat | 17–24 | 25–30 | >32 | >35 | >40 | Flowering & grain filling most sensitive; yield loss above 32–35°C | |
| Rice | 17–24 | 25–30 | >32 | >35 | >40 | Spikelet sterility above 35°C at flowering | |
| Soybean | 17–24 | 25–30 | >32 | >35 | >40 | Pod set and seed fill most sensitive; yield loss above 35°C | |
| Sorghum | 17–24 | 25–30 | >32 | >35 | >40 | More tolerant than maize; flowering still sensitive | |
| Banana | 17–24 | 25–30 | >32 | >35 | >40 | Fruit development sensitive to heat stress | |
| Potato | 17–24 | 25–30 | >32 | >35 | >40 | Tuber initiation sensitive to heat; yield loss above 25–30°C | |
| Cassava | 17–24 | 25–30 | >32 | >35 | >40 | Relatively heat tolerant, but root bulking can be affected above 35°C | |
| Oat | 17–24 | 25–30 | >32 | >35 | >40 | Flowering and grain filling sensitive | |
| Tobacco | 17–24 | 25–30 | >32 | >35 | >40 | Leaf development sensitive to heat stress | |
| Tomato | 17–24 | 25–30 | >32 | >35 | >40 | Fruit set and development sensitive; yield loss above 32–35°C | |
| Fisheries (General) | 17–24 | 25–30 | >32 | >35 | >40 | Species-specific; many fish experience lethal stress above 35–40°C |
Notes:
Humidity: Stress thresholds are lower at higher humidity; short exposures above 35°C (high humidity) or 40°C (low humidity) are often lethal.
Crops: Most field crops are especially vulnerable to heat stress during reproductive stages (flowering, grain/tuber/fruit filling).
Fisheries: Fish species vary, but many follow similar patterns; high temperatures reduce dissolved oxygen and can be lethal.
5 Crop Calendar datasets
The section Crop Calendar datasets provides information on Crop Calendar datasets used to identify average planting (sowing) and maturity (harvest) dates for various major crops. Each dataset offers unique characteristics regarding spatial and the type of crops covered. These datasets include AgMIP-GGCMI Crop Calendar, Jägermeyr Crop Calendar, MIRCA-OS, and WorldCereal Project Crop Calendar.
| Dataset Name | Short Description | Provider/Source | Spatial Coverage | Temporal Coverage | Spatial Resolution | Temporal Resolution | Key Variables | Data Quality Indicators | Metadata Completeness | Update Frequency | Data Provenance | Preprocessing Steps | Access Methods | API/Package Links | Example Scripts | License/Terms | Integration Challenges | Data Ownership | Versioning/Change Log | Data Citation | Benchmarking/Validation Studies | Potential Applications (Agroecological) | Priority Level |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| AgMIP-GGCMI Crop Calendar | Global dataset providing multi-year average planting (sowing) and maturity (harvest) dates for 18 major crops, distinguishing rainfed and irrigated systems, at 0.5° spatial resolution. Composite of multiple observational sources designed for harmonized crop modeling and climate impact assessments. | AgMIP (Agricultural Model Intercomparison and Improvement Project) and GGCMI (Global Gridded Crop Model Intercomparison), led by NASA GISS, Potsdam Institute for Climate Impact Research, University of Minnesota, University of Goettingen. | Global land areas at 0.5° x 0.5° grid cells | Multi-year average (static growing periods; no interannual variability) | 0.5° x 0.5° latitude/longitude | One growing season per crop per grid cell (except wheat and rice with multiple seasons) | Planting day of year (DOY), Maturity (harvest) day of year (DOY) for 18 crops, separately for rainfed and irrigated systems | Gap-filling and spatial extrapolation documented; static averages only; users advised to check regional data quality | Comprehensive metadata including crop-specific details, spatial/temporal coverage, data sources, and processing methods | Static dataset; no regular updates planned | Composite product merging various observational data sources; detailed provenance in Jägermeyr et al. (2021) publication | Spatial extrapolation, gap-filling, merging of observational datasets; no crop rotations included | Download via Zenodo (DOI link); source code and tools on GitHub (link) | No direct API; programmatic access via downloadable scripts and R package on GitHub | R scripts for simulation and adaptation of crop calendars available on GitHub | Open for research and non-commercial use; citation required | Static averages limit temporal dynamics; gap-filled and extrapolated areas require caution; no crop rotation or interannual variability | AgMIP and GGCMI consortium | Version tracked on Zenodo; updates documented in dataset record | Jägermeyr et al. (2021), “Climate impacts on global agriculture emerge earlier in new generation of climate and crop,” Nature Food; DOI: 10.1038/s43016-021-00400-y | Validation and evaluation detailed in Jägermeyr et al. (2021) and Minoli et al. (2022); see dataset documentation | Calibration of crop phenology models; harmonized growing periods for multi-model crop simulations; climate impact assessments; adaptation scenario development | High (core dataset for global crop modeling and climate impact studies) |
| Jägermeyr Crop Calendar | The Jägermeyr Crop Calendar is a global dataset offering multi-year average planting (sowing) and maturity (harvest) dates for 18 major crops, distinguishing rainfed and irrigated systems at 0.5° spatial resolution. It merges multiple observational sources with gap-filling and spatial extrapolation to cover cultivated and uncultivated areas. Designed for global gridded crop model intercomparison and climate impact studies. | Developed by Jonas Jägermeyr and collaborators, hosted by AgMIP-GGCMI with contributions from Potsdam Institute for Climate Impact Research, NASA GISS, and others. | Global land areas at 0.5° x 0.5° grid cells | Multi-year average (static growing periods; no interannual variability) | 0.5° latitude/longitude grid cells | One growing season per crop per grid cell, except wheat (winter and spring) and rice (two main seasons) | Planting day of year (DOY), maturity (harvest) day of year (DOY), separated by rainfed and irrigated management | Gap-filling and spatial extrapolation applied; static averages only; users advised to verify regional data quality | Comprehensive metadata including crop-specific details, spatial/temporal coverage, data sources, and processing methods | Static dataset; no regular updates planned | Composite product merging multiple observational sources; gap-filling and spatial extrapolation for uncultivated or data-sparse areas | Spatial extrapolation, gap-filling, merging multiple observational datasets; no crop rotations included | Download via Zenodo (DOI); source code and modeling scripts on GitHub (link) | No direct API; programmatic access via downloadable scripts and R package on GitHub | R scripts for simulation and adaptation of crop calendars available on GitHub | Open for research and non-commercial use; users should cite original publications appropriately | Static multi-year averages only; no crop rotations or interannual variability; spatial extrapolation may reduce local accuracy | AgMIP-GGCMI consortium | Version tracked on Zenodo; updates documented in dataset record | Jägermeyr et al. (2021), “Climate impacts on global agriculture emerge earlier in new generation of climate and crop,” Nature Food, DOI: 10.1038/s43016-021-00400-y | Validation and evaluation detailed in Jägermeyr et al. (2021) and Minoli et al. (2022); see dataset documentation | Calibration of crop model phenology; harmonized growing periods for multi-model crop simulations; climate impact assessments; adaptation scenario development | High (core dataset for global crop modeling and climate impact studies) |
| MIRCA-OS (Monthly Irrigated and Rainfed Cropped Area, Open Source) | MIRCA-OS is a global, open-source dataset providing monthly irrigated and rainfed cropped area maps for 23 crop classes at 5-arcminute (~10 km) spatial resolution for years 2000, 2005, 2010, and 2015. Combines subnational crop-specific harvested area statistics with global gridded land cover and crop calendar data to produce monthly growing area grids and annual harvested area maps. | Developed by an international team led by researchers associated with AgMIP and global land use modeling groups. | Global (180°W to 180°E longitude, 90°S to 90°N latitude) | Years 2000, 2005, 2010, and 2015 | 5 arcminutes (~10 km x 10 km); annual maps also at 0.5° resolution | Monthly growing area grids representing crop growth stages from planting to maturity | Monthly growing area (ha) for irrigated and rainfed systems, annual harvested area (ha), crop calendars (planting and maturity months) | Areas with no crop presence assigned zero; gap-filling based on regional statistics and land cover; large file sizes require processing tools | Comprehensive metadata including crop classes, spatial/temporal coverage, processing methods | Static dataset for specified years; no continuous updates planned | Combines subnational harvested area statistics with global land cover and crop calendars; downscaled to 5-arcminute grids using area weighting and cropping calendars; gap-filling and mosaicking applied | Data clipped, reprojected; planting and maturity months extracted; downscaling and mosaicking performed; gap-filling applied | Download via HydroShare (link), GitHub (link), and journal supplementary materials | No dedicated API; programmatic access via GitHub scripts and downloadable datasets | Processing scripts available on GitHub repository | Open for research and non-commercial use; citation of original dataset recommended | Large file sizes; requires GIS or scientific programming tools (e.g., Python with xarray, R, QGIS) to process | MIRCA-OS international development team | Version 0.1 tracked on GitHub and HydroShare; updates documented in dataset records | Kebede et al. (2024), “A global open-source dataset of monthly irrigated and rainfed cropped areas (MIRCA-OS) for the 21st century,” Nature Scientific Data, DOI: 10.1038/s41597-024-04313-w | Validation and methodology described in Kebede et al. (2024); see dataset documentation and GitHub | Mapping spatial and temporal distribution of crop areas; differentiating irrigated and rainfed systems; supporting crop phenology and cropping intensity analyses; input for crop models requiring spatially explicit cropped area and calendar information | High (key dataset for spatial-temporal crop area and irrigation studies) |
| WorldCereal Project Crop Calendars | Global, season-specific maps of start (SOS) and end (EOS) of growing seasons for maize and wheat, supporting global crop mapping and monitoring; foundational for generating high-resolution, seasonally updated crop type and irrigation maps. | WorldCereal Consortium (funded by ESA) | Global, stratified into agro-ecological zones (AEZs) | Currently for 2021 (Phase I); designed for seasonal and annual updates | 0.5° x 0.5° latitude/longitude (~50 km at equator); crop type maps at 10 m resolution | Seasonally updated; start and end dates per major growing season per crop | Start of Season (SOS), End of Season (EOS) for maize and wheat (winter and spring cereals); future crops planned | Quality flags and confidence layers provided; simulated data in data-sparse areas; users advised to review confidence | Comprehensive metadata including crop calendars, AEZ stratification, and processing methods | Seasonal updates planned; currently static for 2021 | Merged national/subnational calendars (GEOGLAM, USDA-FAS, FAO, ASAP); Random Forest model trained on ERA5 climate data to estimate SOS/EOS; stratification into zones with similar growing seasons | Data harmonization, spatial modeling, simulation in data gaps; confidence layers generated | Download via figshare (link) and WorldCereal website (link) | No dedicated API; data usable in GIS and remote sensing workflows | Not explicitly stated; open and free under Creative Commons Attribution 4.0 License | Coarser spatial resolution (0.5°) than crop type maps; limited to maize and wheat currently; simulated data in sparse regions requires caution | WorldCereal Consortium and ESA | Version 100 for 2021; updates documented on WorldCereal platforms | Franch et al. (2022), WorldCereal project reports; ESSD Data Descriptor (2023), DOI: 10.5194/essd-15-5491-2023 | Validation and methodology detailed in project publications and reports | Crop-type mapping and monitoring; crop condition and yield forecasting; agroecological zoning; operational monitoring; input for crop models and remote sensing | High (key dataset for global crop seasonality and monitoring) |
Here are detailed summaries and direct excerpts from key sources describing the methodology used to identify planting (sowing) and harvesting (maturity) dates for the major crop calendar datasets:
5.1 AgMIP-GGCMI Crop Calendar and Jägermeyr Crop Calendar (approach for calculation)
Sowing dates are simulated by first classifying the climate seasonality, then applying rules based on either wet season onset (precipitation) or temperature thresholds (temperature), with crop-specific parameters and validation against observed data. This approach enables global, climate-driven estimation of sowing dates for major crops.
5.1.1 Step-by-Step Summary: Determining Sowing Dates
This summary outlines the methodology for simulating sowing dates of major crops under rainfed conditions, as described in the materials and methods section of Waha et al. (2019).
5.1.2 Input Climate Data Preparation
- Gather monthly temperature, precipitation, and number of wet days at 0.5° x 0.5° spatial resolution from the Climatic Research Unit dataset.
- Generate daily mean temperatures by interpolating monthly means.
- Distribute daily precipitation using a weather generator that considers wet/dry day transitions.
5.1.3 Climate Seasonality Classification
- Calculate annual variation coefficients for precipitation (CVprec) and temperature (CVtemp) from past monthly climate data, using an exponential weighted moving average (recent years weighted more heavily).
- Classify each location into one of four seasonality types:
- No temperature and no precipitation seasonality
- Precipitation seasonality
- Temperature seasonality
- Both temperature and precipitation seasonality
- For combined seasonality, check if the coldest month’s mean temperature is above 10°C (no cold season) or ≤10°C (cold season).
5.1.4 Rule-Based Sowing Date Determination by Seasonality Type
- No seasonality: Assign a default sowing date (January 1) for technical reasons.
- Precipitation seasonality:
- Identify the main wet season using the sum of precipitation-to-potential-evapotranspiration (P/PET) ratios over four consecutive months.
- Sowing date is the first wet day (>0.1 mm precipitation) in the main wet season.
- Temperature seasonality:
- Sowing occurs when daily average temperatures exceed a crop-specific threshold (for spring crops) or fall below a threshold (for winter crops needing vernalization).
- The sowing month is when the mean monthly temperature of the past exceeds (or falls below) the threshold; if the last day of the previous month already meets the threshold, that month is chosen.
- Combined seasonality:
- If coldest month >10°C, precipitation determines sowing (as above).
- If coldest month ≤10°C, temperature determines sowing (as above).
5.1.5 Crop-Specific Temperature Thresholds
- Each crop has a globally applied temperature threshold for sowing, based on literature and observed farmer practices.
- For winter crops, the threshold ensures vernalization requirements are met.
5.1.6 Validation
- Simulated sowing dates are compared with observed global data (MIRCA2000).
- For most crops and regions, the difference between simulated and observed sowing dates is less than one month.
Reference:
[1] Waha et al. (2019), “Climate-driven simulation of global crop sowing dates” (PDF)
5.2 General Crop Modeling Methodologies
Thermal time / growing degree days (GDD) approaches are widely used to simulate crop phenology, where maturity is reached once accumulated heat units meet crop-specific thresholds.
For example, the Community Land Model (CLM) uses GDD accumulation to determine maturity dates (Olin et al., 2015;Jagermeyr et al., 2018; Rabin et al, 2023).
Rule-based methods also use climate seasonality, temperature thresholds, and soil moisture indices to estimate sowing and maturity dates (see refs: Mathison et al., 2018; Minoli et al., 2022).
References:
Olin, S., Schurgers, G., Lindeskog, M., Wårlind, D., Smith, B., Bodin, P., Holmér, J., and Arneth, A. (2015). Modelling the response of yields and tissue C : N to changes in atmospheric CO2 and N management in the main wheat regions of western Europe. Biogeosciences, 12, 2489–2515. https://doi.org/10.5194/bg-12-2489-2015
Jägermeyr, J., & Frieler, K. (2018). Spatial variations in crop growing seasons pivotal to reproduce global fluctuations in maize and wheat yields. Science Advances, 4, eaat4517. https://doi.org/10.1126/sciadv.aat4517
Rabin, S. S., Sacks, W. J., Lombardozzi, D. L., Xia, L., & Robock, A. (2023). Observation-based sowing dates and cultivars significantly affect yield and irrigation for some crops in the Community Land Model (CLM5). Geoscientific Model Development, 16, 7253–7273. https://doi.org/10.5194/gmd-16-7253-2023
Mathison, C., Deva, C., Falloon, P., & Challinor, A. J. (2018). Estimating sowing and harvest dates based on the Asian summer monsoon. Earth System Dynamics, 9, 563–592. https://doi.org/10.5194/esd-9-563-2018
Minoli, S., Jägermeyr, J., Asseng, S., Urfels, A., & Müller, C. (2022). Global crop yields can be lifted by timely adaptation of growing periods to climate change. Nature Communications, 13, Article 7079. https://doi.org/10.1038/s41467-022-34411-5
5.3 Other approach
Different approaches are applied in existing crop models to determine current and future sowing dates. Crop models such as LPJmL (Bondeau et al., 2007) identify sowing dates from climate data and crop water and temperature requirements for sowing. The calculation procedure currently applied in LPJmL (Bondeau et al.,2007) is not applicable for all crops in different climatic regions and has only been evaluated for temperate cereals.
Another approach is to optimize sowing dates using the crop model by selecting the date which leads to the highest crop yield,a method applied, for example, in DayCent (Stehfest et al., 2007), or by selecting the optimal growing period based on pre-defined crop-specific requirements, as in GAEZ (Fischer et al.,2002).
Finally, pre-defined sowing dates based on obser vations have been used, e.g. in the Global Crop Water Model (GCWM)(Siebert & Döll,2008) and in GEPIC (Liu et al.,2007).
References:
Bondeau, A., Smith, P. C., Zaehle, S., Schaphoff, S., Lucht, W., Cramer, W., … & Smith, B. (2007). Modelling the role of agriculture for the 20th century global terrestrial carbon balance. Global Change Biology, 13(3), 679-706.
https://gmd.copernicus.org/articles/11/1343/2018/gmd-11-1343-2018.pdfStehfest, E., van Vuuren, D., Kram, T., Bouwman, L., Alkemade, R., Bakkenes, M., … & Prins, A. (2014). Integrated assessment of global environmental change with IMAGE 3.0. Model description and policy applications.
Available via model documentation or scientific databases.Fischer, G., Shah, M., & van Velthuizen, H. (2002). Global Agro-ecological Zones (GAEZ v3.0) — Model documentation.
http://www.fao.org/nr/gaez/en/Siebert, S., & Döll, P. (2008). Quantifying blue and green virtual water contents in global crop production as well as potential production losses without irrigation. Journal of Hydrology, 384(3-4), 198-217.
https://www.sciencedirect.com/science/article/pii/S0022169408003587Liu, J., You, L., Amini, M., Obersteiner, M., Herrero, M., Zehnder, A. J., & Yang, H. (2010). A high-resolution assessment on global nitrogen flows in cropland. Proceedings of the National Academy of Sciences, 107(17), 8035-8040.
https://www.pnas.org/content/107/17/8035
6 Discussion on State-of-the-Art Approaches: Rainfall-Based vs NDVI-Based Seasonality Detection
Seasonality detection plays a vital role in ecological monitoring, agricultural management, and climate studies. Two primary approaches are widely used: rainfall-based and NDVI-based methods, each with distinct data sources, strengths, and limitations.
6.1 Comparison Table of Rainfall-Based and NDVI-Based Seasonality Detection
| Aspect | Rainfall-Based Detection | NDVI-Based Detection |
|---|---|---|
| Data Source | Precipitation measurements (rain gauges, satellites) 1 | Satellite-derived vegetation greenness index (NDVI) 2 |
| Primary Focus | Hydrological inputs and rainfall patterns | Vegetation response and phenology 3 |
| Detection Method | Change-point detection, threshold methods 4 | Time-series analysis of vegetation indices 5 |
| Strengths | Direct measurement of rainfall; timely detection of season transitions 1,4 | Captures biological/ecological response; useful over large spatial scales 2,3 |
| Limitations | Does not capture vegetation or ecological response; may miss lag effects | Influenced by other climatic factors; time lag between rainfall and vegetation response 3,5 |
| Time Lag Consideration | Generally immediate reflection of rainfall events | Vegetation response lags rainfall by weeks to months; lag varies by region and ecosystem 5 |
| Applications | Defining rainy season onset/cessation; hydrological studies 1,4 | Monitoring vegetation dynamics; ecological and agricultural assessments 2,3 |
| Complementarity | Provides physical season boundaries | Validates and refines seasonality based on vegetation growth 2,5 |
6.1.1 Rainfall-Based Seasonality Detection
Rainfall-based approaches utilize precipitation data to identify the onset, duration, and end of rainy seasons through statistical techniques such as change-point detection and threshold analysis. These methods provide direct and timely indicators of hydrological seasonality, which is critical in semi-arid and arid regions where rainfall governs water availability and crop cycles. However, rainfall data alone may not reflect the biological or ecological responses that follow precipitation events 1,4.
6.1.2 NDVI-Based Seasonality Detection
NDVI, derived from satellite remote sensing, measures vegetation greenness and is widely employed to infer seasonality by capturing vegetation phenology. NDVI strongly correlates with rainfall, especially during growing seasons, but the relationship is modulated by time lags, vegetation type, soil properties, and other climatic factors such as temperature and solar radiation. Time lags between rainfall and NDVI response can range from a few weeks to several months and vary spatially, necessitating advanced modeling approaches to accurately capture these dynamics 2,3,5.
6.1.3 Integration and Complementarity
Integrating rainfall and NDVI data provides a more comprehensive view of seasonality by combining physical rainfall patterns with ecological vegetation responses. This integration improves the accuracy of seasonality detection and enhances understanding of ecosystem health and productivity under varying climatic conditions 2,5.
References:
[1] Shisanya, C. A., Recha, C., & Anyamba, A. (2011). Rainfall variability and its impact on normalized difference vegetation index in arid and semi-arid lands of Kenya. International Journal of Geosciences, 2(01), 36. http://dx.doi.org/10.4236/ijg.2011.21004
[2] Dutta, S., Rehman, S., Chatterjee, S., & Sajjad, H. (2021). Analyzing seasonal variation in the vegetation cover using NDVI and rainfall in the dry deciduous forest region of Eastern India. In Forest Resources Resilience and Conflicts (pp. 33-48). Elsevier. https://doi.org/10.1016/B978-0-12-822931-6.00003-4
[3] Chamaille‐Jammes, S., Fritz, H., & Murindagomo, F. (2006). Spatial patterns of the NDVI–rainfall relationship at the seasonal and interannual time scales in an African savanna. International Journal of Remote Sensing, 27(23), 5185-5200. https://doi.org/10.1080/01431160600702392
[4] Guo, J., Liao, W., Qimuge, H., Xu, Y., Wang, J., & Narisu. (2025). Seasonal analysis of spatial and temporal variations in NDVI and its driving factors in Inner Mongolia during the vegetation growing season (1999–2019). Frontiers in Forests and Global Change, 8, 1555385. https://doi.org/10.3389/ffgc.2025.1555385
[5] Wu, T., Feng, F., Lin, Q., & Bai, H. (2019). Advanced method to capture the time-lag effects between annual NDVI and precipitation variation using RNN in the arid and semi-arid grasslands. Water, 11(9), 1789. https://www.osmikon.de/osmikonsearch/Record/baseftdoajarticlesoaidoajorg_article8cb7f18a982d4c5eaade0acd88eed449
7 Prioritization Methodology
7.1 Dataset Prioritization Methodology
To ensure the selection of optimal climate datasets for agroecological research, we apply a structured prioritization approach. This process evaluates datasets based on criteria critical for agricultural applications, data integration, and scientific rigor.
7.1.1 Prioritization Criteria
- Relevance to Agroecology
- Does the dataset include variables essential for crop modeling and agricultural risk assessment (e.g., precipitation, temperature, soil moisture, evapotranspiration)?
- Spatial and Temporal Alignment
- Does the dataset’s spatial and temporal resolution meet the needs of farm-level or regional analysis (e.g., <10 km spatial resolution, daily or sub-monthly time steps)?
- Accessibility
- Is the dataset freely available for research use? Does it offer clear documentation and robust access methods (e.g., API, direct download, cloud platforms)?
- Integration Potential
- Are there established tools, scripts, or packages (e.g., Python/R libraries, Google Earth Engine scripts) to facilitate data retrieval, preprocessing, and analysis?
- Data Quality, Validation, and Uncertainty
- Has the dataset been peer-reviewed, validated against ground observations, or subject to uncertainty quantification? Is metadata comprehensive and transparent?
7.1.2 Example Scoring Approach
| Criterion | Weight (%) | Score (1–5) | Weighted Score |
|---|---|---|---|
| Relevance to Agroecology | 30 | 5 | 1.5 |
| Spatial/Temporal Alignment | 20 | 4 | 0.8 |
| Accessibility | 20 | 3 | 0.6 |
| Integration Potential | 20 | 4 | 0.8 |
| Data Quality & Validation | 10 | 4 | 0.4 |
| Total | 100 | 4.1 |
Priority Level:
Formula:
Weighted Score = (Weight as decimal) × Score
- High: Score ≥ 3.0 (Core dataset for agroecology)
- Medium: 2.0 ≤ Score < 3.0 (Supplementary or niche use)
- Low: Score < 2.0 (Limited utility for project)
7.1.3 Prioritized Dataset List
| Rank | Dataset | Weighted Score | Notes/Strengths |
|---|---|---|---|
| 1 | CHIRPS | 4.8 | Essential precip, high res, strong integration |
| 1 | AgERA5 | 4.8 | Key ag variables, high res, tailored for agriculture |
| 2 | TerraClimate | 4.7 | Water balance, high res, global, accessible |
| 2 | IMERG | 4.7 | High-res precip, global, accessible |
| 2 | TAMSAT | 4.7 | Africa-focused, very high res, accessible |
| 2 | CHIRTS | 4.7 | Essential temp, high res, strong integration |
| 3 | ERA5 | 4.2 | Broad variables, high res, strong integration |
| 4 | NEX-GDDP | 4.1 | Future projections, downscaled, accessible |
| 5 | NASA POWER | 4.0 | Solar, temp, global, but coarser res |
| 6 | CMIP6 | 3.9 | Future scenarios, coarser res, less direct for ag |
7.2 Indicator Prioritization Methodology
Indicators were prioritized using a multi-criteria matrix, scoring each index on relevance to key hazards, scientific consensus, crop/livestock specificity, data availability, computation feasibility, and management impact. Those with the highest total scores were selected as core indicators for our climate data tool. This approach ensures that our indicator set is robust, actionable, and grounded in both science and practical application.
7.2.1 Indicator Prioritization Matrix
We used a multi-criteria matrix to prioritize climate hazard indicators for inclusion in our climate data tool. Each indicator was scored from 1 (lowest) to 3 (highest) against the following criteria:
- Relevance to key agroecological hazards (drought, heat, waterlogging, compound events)
- Scientific consensus and prevalence in literature/operational use
- Crop/livestock specificity (can thresholds be tailored?)
- Data availability (open-access, spatial/temporal coverage)
- Computation feasibility (existing scripts, APIs, packages)
- Impact on management and decision-making
| Indicator | Relevance | Scientific Consensus | Crop/Livestock Specificity | Data Availability | Computation Feasibility | Management Impact | Total Score | Priority |
|---|---|---|---|---|---|---|---|---|
| SPI (Standardized Precipitation Index) | 3 | 3 | 1 | 3 | 3 | 2 | 15 | High |
| SPEI (Standardized Precipitation Evapotranspiration Index) | 3 | 3 | 1 | 3 | 3 | 2 | 15 | High |
| Soil Moisture Anomaly | 3 | 2 | 2 | 2 | 2 | 3 | 14 | High |
| NTX (Number of Hot Days) | 3 | 2 | 2 | 3 | 2 | 3 | 15 | High |
| THI (Temperature Humidity Index, Livestock) | 2 | 2 | 2 | 2 | 2 | 2 | 12 | Medium |
| VPD (Vapor Pressure Deficit) | 2 | 2 | 2 | 2 | 2 | 2 | 12 | Medium |
| Days Soil Moisture > Field Capacity | 2 | 1 | 2 | 2 | 2 | 2 | 11 | Medium |
Each criterion is scored from 1 (lowest) to 3 (highest). The total score is the sum across all criteria.
8 Glossary of Terms
This section serves as the central reference point for essential terms and definitions foundational to our agroecological (AE) project. It provides a clear and accessible glossary to ensure consistent understanding and application of key concepts throughout the project, supporting effective communication and knowledge sharing among all users. Many definitions herein are adapted from ISO standards (ISO 14050). Climate hazard definitions are sourced from a relevant agriculture-focused project’s GitHub repository.
| Terms | Definitions |
|---|---|
| Climate | Statistical description of the weather in terms of the mean and variability of relevant quantities over a period of time ranging from months to thousands or millions of years |
| Climate scenario | Plausible and often simplified representation of the future climate, based on an internally consistent set of climatological relationships that has been constructed for explicit use in investigating the potential consequences of anthropogenic climate change |
| Climate change | Change in climate that persists for an extended period, typically decades or longer |
| Climate change exposure | Potential impact of climate change on locations and their social, economic, and natural structures |
| Climate action | Human intervention to achieve climate change measures or goals based on mitigation or adaptation priorities under climate change policies |
| Indicator | An indicator is a quantitative, qualitative, or binary variable that can be measured, calculated, or described, representing the status of operations, management, conditions, or impacts |
| Hazard | Potential source of injury or damage to the health of people, or damage to property or the environment |
| Site | Location with geographical boundaries, and on which activities under the control of an organization can be carried out |
| Heat stress | … |
| Drought stress | … |
| Waterlogging & flodding stress | … |
References:
Guerrero-Hidalga et al. (2020). Methodology to Prioritize Climate Adaptation Measures in Urban Areas. Barcelona and Bristol Case Studies. Sustainability, 12(12), 4807.
Hinkel, J. (2009). The PIAM Approach: A Framework for Multi-Criteria Selection of Indicators for Integrated Assessment. Ecological Indicators, 9(1), 89–103.
ISO (2020). ISO 14050:2020(en) Environmental management — Vocabulary: Terms relating to climate change and climate action. ISO, 14050:2020(en).
Adaptation Atlas (2023). Hazards-definitions. GitHub repository.