🛰️ Landsat 1972–present 🌍 Sentinel-2 2015–present 🌫️ Sentinel-5P 2018–present 🌧️ CHIRPS 1981–present ⛰️ 6 DEM sources

Turn Earth observation into
research-ready data —
without writing a line of code.

IndexEngine is a batch-processing console for Google Earth Engine: draw an area of interest, pick a sensor and date range, and generate publication-grade indices, time series, change maps, and reports effortlessly.

Capabilities

One console, nine workflows

Configure your AOI and sensor once, generate everything below seamlessly.

🛰️

Satellite Indices

NDVI, EVI, SAVI, NDWI, NBR, and 20+ more — computed with each sensor's correct native bands.

🧮

Custom Index Builder

Write your own formula with band names and standard math. It behaves exactly like a built-in index.

📈

Time Series

Track any index over daily-to-yearly intervals, with trend regression and percent change anomalies.

🔄

Change Detection

Compare two dates with difference, ratio, or a fully configurable three-class classification.

🌫️

Air Quality

NO2, SO2, CO, O3, CH4, HCHO, and Aerosol Index from Sentinel-5P TROPOMI over your date range.

🌧️

Precipitation

Total accumulated rainfall from CHIRPS, summed over any date range from 1981 to present.

⛰️

Terrain Analysis

Elevation, slope, aspect, hillshade, curvature, and contour bands from six global DEM sources.

📄

Publication Reports

One click produces a formatted PDF with your AOI, formulas, statistics, and charts.

💻

Export to GEE JS

Generate a clean, commented Earth Engine script for peer review and perfect reproducibility.

The Workflow

Four steps to output

Draw your AOI

Sketch a polygon directly on the map, or upload a GeoJSON, Shapefile, or KML/KMZ file.

Choose your source

Pick a sensor family and mission, date range, and a cloud-cover threshold.

Select generation targets

Tick any combination of built-in/custom indices, or switch to Time Series/Change Detection.

Export results

Download GeoTIFFs, CSVs, charts, formatted PDF reports, or ready-to-run GEE JS scripts.

Scientific transparency

The Complete Index Reference

Every index — built-in or custom — is documented with its exact formula, what it measures, typical use cases, known limitations, and the original scientific references. Click any entry to expand it.

🌿 Vegetation

NDVI requires: nir, red

NDVI = (NIR − RED) / (NIR + RED)

What it measures: The oldest and most widely used vegetation index. Values range from −1 to 1; healthy dense vegetation is typically 0.6–0.9, sparse vegetation 0.2–0.5, and water/bare soil near 0 or negative.

Typical use cases: General vegetation vigor, biomass, and phenology monitoring; crop health; deforestation/land-cover change over time.

Limitations: Saturates in dense canopy (loses sensitivity above ~LAI 3); sensitive to soil background and atmospheric conditions in uncorrected imagery.

Rouse, J.W. et al. (1974). Monitoring vegetation systems in the Great Plains with ERTS. NASA SP-351, 309–317.

EVI requires: nir, red, blue

EVI = 2.5 × (NIR − RED) / (NIR + 6·RED − 7.5·BLUE + 1)

What it measures: Enhanced Vegetation Index — optimized to reduce canopy background and atmospheric influences using the blue band as a correction term.

Typical use cases: Dense/high-biomass vegetation monitoring where NDVI saturates; large-area vegetation time series (originally designed for MODIS).

Limitations: Coefficients (6, 7.5, gain 2.5) were tuned for MODIS; applying them to Landsat/Sentinel-2 is a common but approximate adaptation. Requires a clean blue band, which is itself more atmosphere-sensitive.

Huete, A. et al. (2002). Overview of the radiometric and biophysical performance of the MODIS vegetation indices. Remote Sensing of Environment, 83(1–2), 195–213.

SAVI requires: nir, red

SAVI = (NIR − RED) / (NIR + RED + L) × (1 + L), L = 0.5

What it measures: Soil-Adjusted Vegetation Index — introduces a fixed soil-brightness correction factor L to reduce soil noise in sparsely vegetated areas.

Typical use cases: Arid/semi-arid regions and early-season crops where soil background strongly influences the signal.

Limitations: L=0.5 is a general-purpose compromise, not scene-optimized; a poorly-matched L can under/over-correct for a given soil type.

Huete, A.R. (1988). A soil-adjusted vegetation index (SAVI). Remote Sensing of Environment, 25(3), 295–309.

MSAVI requires: nir, red

MSAVI = (2·NIR + 1 − √((2·NIR + 1)² − 8(NIR − RED))) / 2

What it measures: Modified SAVI — self-adjusts the soil correction factor per pixel instead of using a fixed L, via an iterative-equivalent closed form.

Typical use cases: Same applications as SAVI without needing to hand-tune L; useful across mixed soil backgrounds within a single AOI.

Limitations: More complex/less intuitive than NDVI; can behave unstably for extreme reflectance ratios (near-zero denominators).

Qi, J. et al. (1994). A modified soil adjusted vegetation index. Remote Sensing of Environment, 48(2), 119–126.

GNDVI requires: nir, green

GNDVI = (NIR − GREEN) / (NIR + GREEN)

What it measures: Green NDVI — substitutes green for red, making it more sensitive to chlorophyll concentration than standard NDVI.

Typical use cases: Nitrogen-status / chlorophyll-content estimation in precision agriculture; later-stage crop monitoring when NDVI saturates.

Limitations: Still saturates at high canopy density; less field-validated than NDVI for general land-cover work.

Gitelson, A.A. et al. (1996). Use of a green channel in remote sensing of global vegetation from EOS-MODIS. Remote Sensing of Environment, 58(3), 289–298.

ARVI requires: nir, red, blue

ARVI = (NIR − (RED − γ(RED − BLUE))) / (NIR + (RED − γ(RED − BLUE))), γ = 1

What it measures: Atmospherically Resistant Vegetation Index — uses the difference between red and blue reflectance to self-correct for aerosol scattering.

Typical use cases: Vegetation monitoring over hazy/smoke-affected or high-aerosol scenes where atmospheric correction is imperfect.

Limitations: Introduces sensitivity to blue-band calibration noise, which is often the noisiest band on a given sensor.

Kaufman, Y.J. & Tanre, D. (1992). Atmospherically resistant vegetation index (ARVI) for EOS-MODIS. IEEE Transactions on Geoscience and Remote Sensing, 30(2), 261–270.

BGI requires: blue, green

BGI = BLUE / GREEN

What it measures: Blue-Green pigment Index — a simple band ratio sensitive to carotenoid/pigment concentration and vegetation stress.

Typical use cases: Early-stress / pigment-change detection, sometimes used alongside NDVI-family indices as a secondary stress indicator.

Limitations: Far less standardized/validated than NDVI-family indices; highly sensor- and illumination-dependent.

Zarco-Tejada, P.J. et al. (2005). Assessing vineyard condition with hyperspectral indices. Remote Sensing of Environment, 99(3), 271–287.

GEMI requires: nir, red

η = (2(NIR² − RED²) + 1.5·NIR + 0.5·RED) / (NIR + RED + 0.5) GEMI = η(1 − 0.25η) − (RED − 0.125) / (1 − RED)

What it measures: Global Environmental Monitoring Index — a non-linear vegetation index designed to minimize soil background influence without requiring a soil-line parameter like SAVI's L.

Typical use cases: Vegetation monitoring across variable soil backgrounds/atmospheric conditions without scene-specific calibration.

Limitations: Non-linear formulation is harder to interpret intuitively than NDVI; can amplify noise at very low reflectance values (RED near 1 causes instability).

Pinty, B. & Verstraete, M.M. (1992). GEMI: a non-linear index to monitor global vegetation from satellites. Vegetatio, 101(1), 15–20.

💧 Water

NDWI requires: green, nir

NDWI = (GREEN − NIR) / (GREEN + NIR)

What it measures: McFeeters' Water Index — the classic open-water delineation index; water strongly absorbs NIR while reflecting green light.

Typical use cases: Open water body mapping (lakes, rivers, reservoirs), flood extent monitoring.

Limitations: Confuses built-up/urban surfaces with water in some scenes (use MNDWI instead in urban areas).

McFeeters, S.K. (1996). The use of the Normalized Difference Water Index (NDWI) in the delineation of open water features. International Journal of Remote Sensing, 17(7), 1425–1432.

MNDWI requires: green, swir1

MNDWI = (GREEN − SWIR1) / (GREEN + SWIR1)

What it measures: Modified NDWI — replaces NIR with SWIR1, substantially reducing built-up/urban false positives compared to NDWI.

Typical use cases: Water mapping in mixed urban-rural landscapes; the generally preferred water index for Landsat/Sentinel-2 work.

Limitations: Can still misclassify deep shadow (mountains, tall buildings) as water.

Xu, H. (2006). Modification of normalised difference water index (NDWI) to enhance open water features. International Journal of Remote Sensing, 27(14), 3025–3033.

NDMI requires: nir, swir1

NDMI = (NIR − SWIR1) / (NIR + SWIR1)

What it measures: Normalized Difference Moisture Index (also called NDII) — sensitive to vegetation and soil water/moisture content via the SWIR band's water-absorption response.

Typical use cases: Drought monitoring, vegetation water-stress assessment, fuel-moisture estimation for fire risk.

Limitations: Influenced by canopy structure and leaf area, not moisture content alone.

Gao, B.C. (1996). NDWI – A normalized difference water index for remote sensing of vegetation liquid water from space. Remote Sensing of Environment, 58(3), 257–266.
Hardisky, M.A. et al. (1983). The influence of soil salinity, growth form, and leaf moisture on the spectral radiance of Spartina alterniflora. Photogrammetric Engineering and Remote Sensing, 49, 77–83.

AWEI requires: blue, green, nir, swir1, swir2

AWEI = BLUE + 2.5·GREEN − 1.5(NIR + SWIR1) − 0.25·SWIR2

What it measures: Automated Water Extraction Index — a linear combination designed to efficiently separate water from non-water including shadow-prone surfaces.

Typical use cases: Automated large-scale water mapping, especially where shadow (from mountains/clouds) causes false positives with other water indices.

Limitations: Coefficients were tuned/validated primarily on Landsat imagery; less validated on other sensors.

Feyisa, G.L. et al. (2014). Automated Water Extraction Index: A new technique for surface water mapping using Landsat imagery. Remote Sensing of Environment, 140, 23–35.

🏙️ Urban / Soil

NDBI requires: swir1, nir

NDBI = (SWIR1 − NIR) / (SWIR1 + NIR)

What it measures: Normalized Difference Built-up Index — impervious/urban surfaces reflect more strongly in SWIR than NIR relative to vegetation.

Typical use cases: Urban/impervious surface mapping, urban expansion and sprawl monitoring.

Limitations: Bare soil often has a similar spectral signature to built-up surfaces, causing confusion between the two classes.

Zha, Y. et al. (2003). Use of normalized difference built-up index in automatically mapping urban areas from TM imagery. International Journal of Remote Sensing, 24(3), 583–594.

BSI requires: swir1, red, nir, blue

BSI = ((SWIR1 + RED) − (NIR + BLUE)) / ((SWIR1 + RED) + (NIR + BLUE))

What it measures: Bare Soil Index — combines both visible and infrared bands to separate exposed/bare soil from vegetated and built-up surfaces.

Typical use cases: Soil exposure mapping, land degradation and desertification monitoring, complementing NDBI to disambiguate bare soil from built-up.

Limitations: Mixed pixels (partial vegetation cover) and seasonal soil-moisture variation both affect the signal.

Rikimaru, A. et al. (2002). Tropical forest cover density mapping. Tropical Ecology, 43(1), 39–47.

IBI requires: swir1, nir, green, red

IBI = (NDBI − avg(SAVI, MNDWI)) / (NDBI + avg(SAVI, MNDWI))

What it measures: Index-based Built-up Index — jointly combines NDBI, SAVI, and MNDWI into a single ratio, explicitly suppressing the bare-soil and water false positives that affect NDBI when used alone.

Typical use cases: Impervious-surface / built-up-area mapping where higher accuracy than NDBI alone is needed, e.g. urban footprint delineation and imperviousness change detection.

Limitations: Requires four bands (green, red, nir, swir1), so it is unavailable on Landsat MSS; still subject to mixed-pixel confusion at the urban/bare-soil edge, and its three-index composition makes it more sensitive to cloud/shadow contamination in any one input band.

Xu, H. (2008). A new index for delineating built-up land features in satellite imagery. International Journal of Remote Sensing, 29(14), 4269–4276.

ALBEDO requires: blue, red, nir, swir1, swir2

α = (0.356·BLUE + 0.130·RED + 0.373·NIR + 0.085·SWIR1 + 0.072·SWIR2 − 0.0018) / 1.016

What it measures: Broadband surface albedo — a narrowband-to-broadband conversion estimating the fraction of incoming solar radiation reflected by the surface.

Typical use cases: Surface energy balance studies, urban heat island research, land-surface climate modeling inputs.

Limitations: This is an approximation calibrated originally for Landsat ETM+ narrowbands; applying it to other sensors (Sentinel-2, TM, MSS) introduces additional uncertainty. Not a lab-calibrated in-situ measurement.

Liang, S. (2001). Narrowband to broadband conversions of land surface albedo I: Algorithms. Remote Sensing of Environment, 76(2), 213–238.

🌡️ Thermal

LST requires: thermal

LST (°C) = (ST_Bxx × 0.00341802 + 149.0) − 273.15

What it measures: Land Surface Temperature, taken directly from the USGS Collection-2 Level-2 Surface Temperature product (already atmospherically corrected and emissivity-adjusted upstream by USGS), then converted from Kelvin to Celsius.

Typical use cases: Urban heat island analysis, thermal comfort studies, drought/evapotranspiration proxies, wildfire hotspot detection.

Limitations: Only valid for clear-sky pixels (cloud-masked elsewhere); emissivity assumptions are baked into the USGS product, not user-adjustable here; not available for Sentinel-2 or Landsat MSS, which carry no thermal sensor.

USGS (2020). Landsat Collection 2 Level-2 Science Product Guide, Version 4.0.

🧪 Water Quality

NDTI requires: red, green

NDTI = (RED − GREEN) / (RED + GREEN)

What it measures: Normalized Difference Turbidity Index — an optical proxy where higher values indicate more suspended sediment / turbid water.

Typical use cases: Relative turbidity mapping and change-detection in rivers, reservoirs, and coastal waters.

Limitations: Uncalibrated proxy, not a measured NTU/FNU value; relationship between NDTI and true turbidity is scene- and sensor-dependent and needs local field validation.

Lacaux, J.P. et al. (2007). Classification of ponds from high-spatial resolution remote sensing: Application to Rift Valley Fever epidemics in Senegal. Remote Sensing of Environment, 106(1), 66–74.

NDCI requires: rededge1, red

NDCI = (REDEDGE1 − RED) / (REDEDGE1 + RED)

What it measures: Normalized Difference Chlorophyll Index — uses the red-edge band's sensitivity to chlorophyll-a absorption/fluorescence to detect algal presence.

Typical use cases: Algal bloom detection and relative chlorophyll-a mapping in inland/coastal waters (requires a red-edge band, so Sentinel-2 only).

Limitations: Originally developed for MERIS's dedicated red-edge/NIR bands; adapting it to Sentinel-2's Band 5 is a reasonable but imperfect substitution.

Mishra, S. & Mishra, D.R. (2012). Normalized difference chlorophyll index: A novel model for remote estimation of chlorophyll-a concentration in turbid productive waters. Remote Sensing of Environment, 117, 394–406.

CHLA_HIGH requires: rededge1, red

Chl-a = 19.866 × (REDEDGE1 / RED)^2.3051

What it measures: Empirical chlorophyll-a retrieval model optimized for HIGH-concentration algal blooms in inland reservoirs, calibrated against in-situ Sentinel-2 matchups.

Typical use cases: Quantitative Chl-a estimation (µg/L) during bloom events in eutrophic inland waterbodies.

Limitations: Calibrated on specific Mediterranean/Spanish reservoirs — treat as a relative indicator and validate locally with water samples before using absolute values elsewhere.

Soria-Perpinyà, X. et al. (2021). Validation of Sentinel-2 for chlorophyll-a retrieval in Mediterranean reservoirs. Science of the Total Environment, 786, 147536.

CHLA_LOW requires: green, blue, red

Chl-a = 10^(−2.4792 × log₁₀(max(GREEN, BLUE) / RED) + 3.47941)

What it measures: Companion empirical model optimized for LOW-concentration chlorophyll-a in clearer inland waters, using a blue/green-to-red band-ratio algorithm (structurally similar to classic ocean-color OC-type algorithms).

Typical use cases: Chl-a estimation in mesotrophic/oligotrophic lakes and reservoirs where CHLA_HIGH would be less accurate.

Limitations: Same regional-calibration caveat as CHLA_HIGH; band-ratio ocean-color algorithms are also known to be less reliable in optically-complex (high-sediment) inland waters.

Soria-Perpinyà, X. et al. (2021). Validation of Sentinel-2 for chlorophyll-a retrieval in Mediterranean reservoirs. Science of the Total Environment, 786, 147536.

CWQI requires: red

CWQI = −163.31 × RED + 111.43

What it measures: A simple single-band empirical linear regression relating red-band reflectance to a Comprehensive Water Quality Index, calibrated on a specific river system (commonly cited pattern from Weihe River, China studies).

Typical use cases: Included as a worked example of how to build a targeted local regression once you have field water-quality samples paired with satellite reflectance.

Limitations: The coefficients are region- and sensor-specific. They will almost certainly be inaccurate for any other river, lake, or reservoir unless recalibrated against local field measurements.

Representative of empirical regression approaches common in regional Chinese river water-quality remote-sensing studies (e.g., Weihe River basin); recalibrate with local data before use.

NH4N requires: red

NH₄⁺-N = 11.17 × RED − 0.46

What it measures: Empirical linear regression estimating ammonia-nitrogen concentration from red-band reflectance, same regional-calibration pattern as CWQI.

Typical use cases: Template for building a local ammonia-nitrogen regression once field samples are available.

Limitations: Not transferable outside its original calibration region/sensor without recalibration.

Representative of empirical regression approaches in regional river water-quality remote sensing (e.g., Weihe River basin); recalibrate with local data before use.

TP requires: red

TP = 0.64 × RED − 0.027

What it measures: Empirical linear regression estimating total phosphorus concentration from red-band reflectance, same regional-calibration pattern as CWQI/NH4N.

Typical use cases: Template for building a local total-phosphorus regression once field samples are available.

Limitations: Not transferable outside its original calibration region/sensor without recalibration.

Representative of empirical regression approaches in regional river water-quality remote sensing (e.g., Weihe River basin); recalibrate with local data before use.

🔥 Fire / Snow

NBR requires: nir, swir2

NBR = (NIR − SWIR2) / (NIR + SWIR2)

What it measures: Normalized Burn Ratio — the standard index for assessing fire effects; healthy vegetation has high NIR/low SWIR2, burned areas the reverse.

Typical use cases: Burn-severity mapping via dNBR (pre-fire NBR minus post-fire NBR); post-fire vegetation recovery monitoring.

Limitations: A single-date NBR only shows relative vegetation/moisture condition — true burn-severity classification needs a pre-fire/post-fire pair (dNBR).

Key, C.H. & Benson, N.C. (2006). Landscape Assessment: Ground measure of severity, the Composite Burn Index; and Remote sensing of severity, the Normalized Burn Ratio. USDA Forest Service FIREMON Report.

BAI requires: red, nir

BAI = 1 / ((0.1 − RED)² + (0.06 − NIR)²)

What it measures: Burn Area Index — highlights charcoal/burned surfaces by measuring distance in red-NIR spectral space from a reference char-signature point.

Typical use cases: Rapid visual highlighting of recently burned areas, particularly in Mediterranean-type ecosystems where it was developed.

Limitations: Certain non-burned surfaces (some bare soils, dark urban materials) can spectrally resemble char and cause false positives; less reliable outside the ecosystems it was validated in.

Martin, M.P. (1998). Cartografia e inventario de incendios forestales en la Peninsula Iberica a partir de imagenes NOAA-AVHRR. PhD Thesis, Universidad de Alcala.
Chuvieco, E. et al. (2002). Design of an empirical index to estimate burn severity: the Burned Area Index. International Journal of Remote Sensing, 23(23), 5103–5110.

NDSI requires: green, swir1

NDSI = (GREEN − SWIR1) / (GREEN + SWIR1)

What it measures: Normalized Difference Snow Index — snow is highly reflective in visible/green but strongly absorbing in SWIR, the opposite of clouds.

Typical use cases: Snow cover mapping and monitoring, snow/cloud discrimination (the basis of the operational MODIS snow-cover product).

Limitations: Can be confused by ice on lakes/rivers; thin or patchy snow under vegetation canopy is often underestimated.

Hall, D.K. et al. (1995). Development of methods for mapping global snow cover using moderate resolution imaging spectroradiometer data. Remote Sensing of Environment, 54(2), 127–140.

⛰️ Terrain Products

ELEVATION

z = DEM(x, y)

What it measures: Raw terrain height above a reference datum (typically EGM96/WGS84 ellipsoid, depending on the DEM source), in meters.

Typical use cases: Basic terrain context, watershed delineation inputs, siting/suitability analysis.

Limitations: Vertical accuracy and voids vary by source (SRTM has known data voids in steep/mountainous terrain that NASADEM partly fills).

Farr, T.G. et al. (2007). The Shuttle Radar Topography Mission. Reviews of Geophysics, 45(2).

SLOPE

Slope = arctan(√((∂z/∂x)² + (∂z/∂y)²))

What it measures: Steepness of terrain in degrees, computed from the DEM's local gradient.

Typical use cases: Landslide susceptibility mapping, erosion risk, construction/infrastructure siting, hydrological modeling.

Limitations: Sensitive to DEM resolution and vertical noise; coarser DEMs smooth out true local slope extremes.

Zevenbergen, L.W. & Thorne, C.R. (1987). Quantitative analysis of land surface topography. Earth Surface Processes and Landforms, 12(1), 47–56.

ASPECT

Aspect = arctan2(∂z/∂y, −∂z/∂x)

What it measures: Compass direction (0–360°) that a slope faces, derived from the DEM's gradient direction.

Typical use cases: Solar radiation/insolation modeling, microclimate and vegetation-distribution studies, avalanche/landslide directional risk.

Limitations: Undefined/noisy on perfectly flat terrain; same resolution sensitivity as slope.

Zevenbergen, L.W. & Thorne, C.R. (1987). Quantitative analysis of land surface topography. Earth Surface Processes and Landforms, 12(1), 47–56.

HILLSHADE

Hillshade = f(Slope, Aspect, Sun Azimuth, Sun Elevation)

What it measures: Simulated illumination of the terrain surface under a given sun position, producing a shaded-relief visualization.

Typical use cases: Cartographic visualization, visually validating terrain features and drainage patterns.

Limitations: A visualization aid, not an analytical/quantitative product.

Standard GIS terrain-analysis technique; see Burrough, P.A. & McDonnell, R.A. (1998). Principles of Geographical Information Systems. Oxford University Press.

CURVATURE

Curvature ≈ ∇²z (Laplacian filter over the DEM)

What it measures: Rate of change of slope — positive values indicate convex (ridge-like) terrain, negative values indicate concave (valley-like) terrain.

Typical use cases: Landform classification, erosion/deposition zone identification, hydrological flow-accumulation analysis.

Limitations: A Laplacian-kernel approximation used here as a practical proxy for true profile/plan curvature; highly sensitive to DEM noise, benefits from a smoothed DEM input.

Moore, I.D. et al. (1991). Digital terrain modelling: A review of hydrological, geomorphological, and biological applications. Hydrological Processes, 5(1), 3–30.

CONTOUR_BANDS

CONTOUR_BANDS = ⌊z / interval⌋ × interval

What it measures: Elevation classified into discrete bands at a user-chosen interval — a raster approximation of a contour map.

Typical use cases: Quick visual terrain stratification when true vector contour lines aren't required.

Limitations: Not a substitute for true vector contour lines (which GeoTIFF cannot store); band edges are blocky raster boundaries, not smooth contour curves.

Standard raster-reclassification technique; no single primary reference.

🌫️ Air Quality (Sentinel-5P TROPOMI)

NO2 — Nitrogen Dioxide mol/m²

Tropospheric NO₂ column density (mol/m²)

What it measures: TROPOMI-derived tropospheric NO2 vertical column density, a marker of combustion emissions (traffic, power generation, industry).

Typical use cases: Urban air-pollution monitoring, traffic/industrial emission hotspot detection, policy-impact studies (e.g. lockdown NO2 drops).

Limitations: Column density, not ground-level concentration; strongly affected by cloud cover and requires temporal averaging for a stable signal.