Tools for researchers
Free tools treat the soil horizon by horizon. Our advanced methods treat it as one connected system, continuous with depth and across the landscape, because that is how water, carbon and energy actually move.
These tools follow the processes behind the environment: how landscapes form and how water, heat and carbon move through them. They use published, process-based mathematics, so every result can be traced back to its formula and checked.
Every tool runs on your own computer, in your browser. Files you open here are never sent to us or stored anywhere. We can't see them, even if we wanted to. That makes the tools safe for confidential data, unpublished results and manuscripts under review.
Don't take our word for it: once this page has loaded, disconnect from the internet. Every tool still works.
Checking your own or a published result? All the plausibility checks are listed on one page →
Profile harmoniser
Fit a physics-informed spline (phySpline) to irregular horizon data and harmonise it to standard depth intervals. Works for continuous properties (carbon, pH, clay) and for ordered classes such as texture.
1. Your data
Columns: profile id, top depth (cm), bottom depth (cm), then one or more property columns. Comma, semicolon or tab separated. Up to 20,000 rows.
2. Settings
Profile view
Harmonised values
Intervals that extend past the bottom of a profile are averaged over the soil that is present and shown in amber. Blank means no soil in that interval.
Our advanced profile method treats the soil as one continuous system: much faster on large datasets, with uncertainty at every depth and boundary-aware profiles linked to water and heat flux.
Get in touch →Method: Flynn, T., Rasaei, Z., Kostecki, R. (2026). Science of the Total Environment. doi:10.1016/j.scitotenv.2026.181621. Open R implementation: rafikisol package.
Soil carbon stocks
Calculate organic carbon stocks for many profiles at once, by summing horizons to fixed depths, with an optional equivalent soil mass basis for comparing sites or sampling dates.
1. Your data
Columns: id, top (cm), bottom (cm), soc (%), bd (bulk density, g/cm³), and optionally cf (coarse fragments, % by volume).
2. Settings
Equivalent soil mass: stock in the soil needed to reach the reference mass, regardless of depth. Useful when bulk density differs between sites or dates.
Stocks (Mg C/ha)
Amber values: the profile is shallower than the reporting depth, so the stock covers only the soil that was sampled.
Can the change be detected?
Use it when planning a monitoring design, before claiming a change in your own results, or before trusting (or paying for) someone else's: check whether the sampling could detect a change that size. This gives the smallest change a design can reliably detect, and how many samples a claimed change would need.
Field CVs for soil carbon stocks are typically 20–50%. Resampling the same locations removes much of the spatial variation from the comparison; the correlation says how much (0.5–0.8 is common).
Garten, C.T., Wullschleger, S.D. (1999). Soil carbon inventories under a bioenergy crop (switchgrass): measurement limitations. J. Environ. Qual. 28, 1359–1365. doi:10.2134/jeq1999.00472425002800040041x
Summing horizons assumes each layer is uniform. Our published work shows this can substantially overestimate carbon stocks compared with assessing the profile at every point. For baselines, credits or reporting, we calculate stocks from depth-continuous profiles with uncertainty.
Flynn, T., Kostecki, R., Rebi, A., Raza, T. (2024). Accessing global soil raster images and equal-area splines to estimate soil organic carbon stocks on the regional scale. Pedosphere. doi:10.1016/j.pedsph.2024.07.004
Get in touch →Stocks: SOC × bulk density × thickness × (1 − coarse fragments), following Poeplau, C., Vos, C., Don, A. (2017). SOIL 3, 61–66. doi:10.5194/soil-3-61-2017 Equivalent soil mass: Ellert, B.H., Bettany, J.R. (1995). Can. J. Soil Sci. 75, 529–538. doi:10.4141/cjss95-075
Water retention: compare pedotransfer functions
Estimate the water retention curve, field capacity, wilting point and available water from texture, carbon and bulk density with six published pedotransfer functions at once. They often disagree, and the spread between them is an honest measure of how uncertain an estimate from texture alone is. Add measured points to see which one fits your soil.
Your soil
Sand 50–2000 µm, silt 2–50 µm, clay < 2 µm. Organic matter is taken as 1.724 × organic carbon, as the original equations assume. Without a bulk density, the Saxton & Rawls estimate for a normally packed soil is used. CEC and pH are needed only for the tropical function (Hodnett & Tomasella). 10 kPa is often a better field capacity for sandy soils.
Your measurements (optional)
Two columns: suction and volumetric water content, one measurement per row, for the soil described on the left. Each function is then scored by how closely it matches your points.
Retention curves
We build depth-continuous retention profiles and calibrate pedotransfer functions to your own site data, ready for water balance and recharge modelling.
Get in touch →Saxton, K.E., Rawls, W.J. (2006). Soil water characteristic estimates by texture and organic matter for hydrologic solutions. Soil Sci. Soc. Am. J. 70, 1569–1578. doi:10.2136/sssaj2005.0117
Rawls, W.J., Brakensiek, D.L. (1985). Prediction of soil water properties for hydrologic modeling. In: Jones, E.B., Ward, T.J. (eds), Watershed Management in the Eighties. ASCE, New York, 293–299.
Wösten, J.H.M., Lilly, A., Nemes, A., Le Bas, C. (1999). Development and use of a database of hydraulic properties of European soils. Geoderma 90, 169–185. doi:10.1016/S0016-7061(98)00132-3
Vereecken, H., Maes, J., Feyen, J., Darius, P. (1989). Estimating the soil moisture retention characteristic from texture, bulk density, and carbon content. Soil Science 148, 389–403.
Hodnett, M.G., Tomasella, J. (2002). Marked differences between van Genuchten soil water-retention parameters for temperate and tropical soils: a new water-retention pedo-transfer functions developed for tropical soils. Geoderma 108, 155–180. doi:10.1016/S0016-7061(02)00105-2
Cosby, B.J., Hornberger, G.M., Clapp, R.B., Ginn, T.R. (1984). A statistical exploration of the relationships of soil moisture characteristics to the physical properties of soils. Water Resour. Res. 20, 682–690. doi:10.1029/WR020i006p00682
van Genuchten curves
Explore how the van Genuchten–Mualem parameters shape water retention and unsaturated conductivity. Start from a texture class, convert any of six pedotransfer functions to van Genuchten form, or fit the curve to your own measurements, then adjust.
Start from
Two columns: suction and volumetric water content. Use 0 for saturation. Fitting four parameters needs at least five points spread from wet to dry; with fewer, or with no points drier than about 1000 kPa, fix θr at 0.
Water retention θ(h)
Hydraulic conductivity K(h)
We fit hydraulic parameters to your lab or field measurements and carry them into two-dimensional flux, recharge and drainage modelling across hillslopes.
Get in touch →van Genuchten, M.Th. (1980). A closed-form equation for predicting the hydraulic conductivity of unsaturated soils. Soil Sci. Soc. Am. J. 44, 892–898. doi:10.2136/sssaj1980.03615995004400050002x Mualem, Y. (1976). A new model for predicting the hydraulic conductivity of unsaturated porous media. Water Resour. Res. 12, 513–522. doi:10.1029/WR012i003p00513 Class parameters: Carsel, R.F., Parrish, R.S. (1988). Developing joint probability distributions of soil water retention characteristics. Water Resour. Res. 24, 755–769. doi:10.1029/WR024i005p00755 Pedotransfer functions: see the Water retention tab. Fitting: least squares on θ, as in RETC (van Genuchten, Leij & Yates 1991).
Hydrologic soil group and runoff
Find a soil's hydrologic soil group (A–D) with the NRCS rules, then see what it means: the runoff curve number for your land cover, including arid and semiarid rangeland, and how much of a storm runs off on dry, average or wet soil.
1. The soil
An impermeable layer is bedrock, a hardpan (fragipan, duripan, petrocalcic or other cemented layer), permafrost, or any layer with Ksat below 0.01 µm/s. Fractured or deeply weathered rock that still transmits water is not impermeable; judge it by its Ksat. A seasonal high water table counts if it occurs in any month of the year. Soil surveys list both for each map unit.
Hydrologic soil group
2. What it means for runoff
We map hydrologic soil groups continuously from geology, terrain and soil data, including where soil surveys are missing or coarse.
Get in touch →Hydrologic soil groups: USDA-NRCS (2009). National Engineering Handbook, Part 630 Hydrology, Chapter 7: Hydrologic Soil Groups (Table 7-1). Curve numbers and runoff equation: USDA-SCS (1986). Urban Hydrology for Small Watersheds, Technical Release 55 (TR-55), Tables 2-2a to 2-2d, average antecedent moisture, initial abstraction 0.2S. Antecedent moisture (AMC I and III): Chow, V.T., Maidment, D.R., Mays, L.W. (1988). Applied Hydrology. McGraw-Hill, New York, Section 5.5. Initial abstraction 0.05S: Hawkins, R.H., Jiang, R., Woodward, D.E., Hjelmfelt, A.T., Van Mullem, J.A. (2002). Runoff curve number method: examination of the initial abstraction ratio. Proceedings of the Second Federal Interagency Hydrologic Modeling Conference, Las Vegas, Nevada. Ksat from texture: Saxton, K.E., Rawls, W.J. (2006). Soil Sci. Soc. Am. J. 70, 1569–1578. doi:10.2136/sssaj2005.0117. This is a screening tool, not a replacement for official soil survey data.
Texture class
Classify particle-size fractions on the USDA, European or Australian texture triangle, one sample at a time or a whole file at once, then download the triangle or the classified data.
One sample
Many samples
Columns must be named sand, silt and clay (in %, any order). An id column is optional and any other columns are kept. Rows that don't add to 100% are rescaled and flagged.
Classified samples
We produce continuous texture surfaces with depth and across landscapes, with uncertainty, and convert between particle-size systems.
Get in touch →Class boundaries: USDA-NRCS Soil Survey Manual; HYPRES / European Soil Database texture classes; Australian triangle as implemented in the soiltexture R package (Moeys; coordinates from B. Minasny). Fractions must be measured on the particle-size limits of the chosen system.
Munsell soil colour
See a Munsell notation as an on-screen colour, browse a soil chart page, learn what the colour usually tells you about the soil, convert colour-sensor readings (L*a*b*, RGB or XYZ) to Munsell with the probability of each nearby chip, one at a time or a whole file, and estimate the colour from a photo with a reference card.
Notation
Chart page
Rows are value (lightness), columns are chroma (strength). Tap a chip to select it.
From a colour sensor or colorimeter
Instruments such as colorimeters, spectrophotometers and pocket colour sensors report colour as three numbers, usually CIELAB (L*, a*, b*) or RGB. This finds the nearest chip on the soil and gley chart pages, how close it is, and the Munsell value calculated directly from lightness.
The repeat-reading spread is how much readings of the same sample differ: about 0.5 for a laboratory spectrophotometer, 1–2 for a pocket sensor, more if the sample surface or moisture varies. It turns the nearest chip into a probability over neighbouring chips. Set it to 0 to see only the nearest chip.
Many samples
Columns named L, a, b (or L*, a*, b*), R, G, B, X, Y, Z, or hex, plus an optional id column. The illuminant setting above applies to L*a*b* and XYZ. RGB and hex are taken as standard sRGB. RGB read from an ordinary photo is not calibrated: the camera's white balance and exposure shift it, so use the photo tool below instead.
From a photo
A photo of a soil face or a ped can give a rough Munsell colour if it includes a reference card. The Soil Colour Reader corrects the photo with an 18% grey card or a black-and-white card, drops shadows, pores and glare, and gives the most likely chip with probabilities over the neighbouring chips. It runs on your device like the other tools.
Open the Soil Colour Reader →Colours are computed from the Munsell renotation data (illuminant C) and converted to sRGB. Screens vary, so use this for learning, not in place of a physical chart. Chips that fall outside the sRGB range are left out. Matching uses the CIEDE2000 colour difference against the chips on the soil and gley pages; Munsell value from luminance follows ASTM D1535. A ΔE₀₀ under about 2 is hard to tell apart by eye; above 5, no chip is a good match. Instrument readings depend on moisture, grinding and the light source, so record the conditions as you would for a chart reading.
Sampling density
Estimate how many samples you need to measure a field or site mean within a chosen margin of error, assuming simple random sampling.
Inputs
No CV to hand? Typical values: pH 5–15%, organic carbon 20–40%, available P or K 40–80%.
Result
Sampling for a map: using the variogram
The calculator above sizes a sample to estimate one average. For a map, what matters is how far apart samples are compared with the variogram range. This finds how many samples on a regular grid give a chosen kriging error, from a variogram you already have (for example from the variogram tool or a nearby study).
Enter the variogram the way it was reported: as the nugget : sill ratio, as the nugget and sill, or as the nugget and partial sill (the sill minus the nugget; gstat reports this as psill, and many papers do too). The sill is the total: nugget + partial sill. Only the ratio matters here, so any of the three gives the same answer.
Map error is the root mean kriging error at unsampled locations, as a share of the standard deviation of the property. 100% means no better than the mean; the nugget sets the lowest error any density can reach.
Burgess, T.M., Webster, R., McBratney, A.B. (1981). Optimal interpolation and isarithmic mapping of soil properties. IV. Sampling strategy. J. Soil Sci. 32, 643–659. doi:10.1111/j.1365-2389.1981.tb01737.x McBratney, A.B., Webster, R. (1983). How many observations are needed for regional estimation of soil properties? Soil Science 135, 177–183.
Random sampling ignores the fact that nearby soils are alike. We design optimised and stratified sampling schemes and management zones that reach the same precision with fewer samples.
Get in touch →Sample size for a mean under simple random sampling, n = (t × CV ÷ E)²: Cochran, W.G. (1977). Sampling Techniques, 3rd ed. Wiley.
Variogram & kriging
Fit a variogram to point data and test ordinary kriging by leave-one-out cross-validation: each point is predicted from its neighbours with itself left out, then compared with what was measured.
1. Your points
Columns: an optional id, coordinates named x and y (projected, e.g. metres) or lon and lat (decimal degrees, converted to km), and one or more numeric properties. Up to 5,000 points.
2. Settings
Log suits skewed, positive data (organic carbon, available P, EC). The variogram and kriging run on the logs; predictions are back-transformed to the original units with the lognormal correction, so they are unbiased means, not medians.
Area under the variogram
Integrates the fitted model between two lag distances, to compare spatial structure over a chosen range of distances, for example short-range against long-range, or the same interval across sites, depths or studies.
Variogram
Cross-validation: observed vs predicted
Where the errors are
Clusters of the same colour suggest a trend or a boundary the variogram doesn't capture. A single large dot is worth checking for a data error.
Cross-validation results
Check a result: yours or a published one
Use it before you submit, or when reviewing. Kriging accuracy is set by the variogram and by how far apart the samples are, so a reported error can be checked without the data. This lays out samples at the stated density, works out the kriging variance at each one from the stated variogram, and gives the cross-validation RMSE you should expect if that variogram is right.
Enter the variogram the way it was reported: as the nugget : sill ratio, as the nugget and sill, or as the nugget and partial sill (the sill minus the nugget; gstat reports this as psill, and many papers do too). The sill is the total: nugget + partial sill. With the ratio, the sill is taken as the variance of the data (SD²); with a nugget and sill, the stated sill is used, and the standard deviation becomes optional.
If the nugget isn't reported, leave it at 0: that is the most generous assumption, so a result that still looks too good really is.
Kriging variance from the variogram and sample layout alone: Burgess, T.M., Webster, R., McBratney, A.B. (1981). J. Soil Sci. 32, 643–659. doi:10.1111/j.1365-2389.1981.tb01737.x Webster, R., Oliver, M.A. (2007). Geostatistics for Environmental Scientists, 2nd ed. Wiley.
Reading the results
- Nugget : sill is the share of variation you can't predict at your sampling scale: measurement error plus variation over distances shorter than your closest samples. Under 0.25 is strong spatial dependence, 0.25–0.75 moderate, over 0.75 weak.
- Range is the distance beyond which samples stop telling you about each other. If your samples are mostly further apart than the range, kriging can do little better than the mean, however good the software.
- RMSE vs. the mean is the real test: it shows how much spatial prediction actually adds. A map made by kriging is only worth its colours if this is clearly above zero.
- MSDR and interval coverage show whether the uncertainty is honest. MSDR near 1 and coverage near 90% mean the kriging variance can be trusted; otherwise the uncertainty map is too optimistic or too cautious.
- Two limits that are rarely stated. Cross-validated RMSE can't fall much below the square root of the nugget, because that part of the variation isn't predictable from neighbours. And when samples are further apart than the range, RMSE approaches the standard deviation of the data. A published RMSE well below these limits usually means the validation points weren't independent of the calibration data, the variogram came from a different dataset, or the error was reported on a different scale.
We produce kriged and regression-kriged maps with terrain, remote sensing and legacy data, full prediction uncertainty, and sampling designs built from the variogram.
Get in touch →Method: classical (Matheron) or robust (Cressie & Hawkins 1980) variogram estimator; model fitted by weighted least squares with weights Nj/hj² (as in gstat); ordinary kriging with the nearest neighbours; Matérn model with fitted smoothness in the standard and Stein (1999) parameterisations; ranges are practical ranges (95% of the sill) for every model; lognormal ordinary kriging back-transform after Journel (1980) and Webster & Oliver (2007); nugget/sill classes after Cambardella et al. (1994).
Model evaluation
Check how well a soil map or model matches independent observations: accuracy and agreement for continuous predictions, prediction-interval coverage, class accuracy, and the confusion index for class probabilities.
Observed and predicted values
Columns named obs and pred. Add sd (prediction standard deviation) for a coverage curve, or lower and upper for the coverage of your own intervals.
Observed vs predicted
Prediction interval coverage (PICP)
Points below the line mean the intervals are too narrow (overconfident); above means too wide.
Check a result: yours or a published one
R² and RMSE constrain each other, and both are limited by how precisely the soil was measured. Enter your own figures before you submit, or the figures from a paper or report, to see whether the numbers can all be true.
SD of the observed values is often in a summary-statistics table. Repeatability is the SD of repeated measurements of the same sample; labs or method papers report it.
No paper needed: for the same validation data, RMSE ≥ SD × √(1 − R²) follows from the definitions of R² and RMSE, and no model can predict more precisely than the soil was measured.
Observed and predicted classes
Columns named obs and pred, one site per row.
Confusion matrix
Rows are observed classes, columns are predicted. Producer's accuracy: share of each observed class predicted correctly. User's accuracy: share of each predicted class that is correct.
Check an accuracy: yours or a published one
An accuracy from a small validation set is uncertain, and it only means something compared with what you would get by guessing. Enter your own figures or a paper's to see the confidence interval and whether the map beats chance or another map.
The largest class share is the accuracy of a "map" that predicts the most common class everywhere. If it isn't given, an even split between classes is assumed, which flatters the map.
Wilson, E.B. (1927). Probable inference, the law of succession, and statistical inference. J. Am. Stat. Assoc. 22, 209–212. doi:10.1080/01621459.1927.10502953 Foody, G.M. (2009). Classification accuracy comparison: hypothesis tests and the use of confidence intervals in evaluations of difference, equivalence and non-inferiority. Remote Sens. Environ. 113, 1658–1663. doi:10.1016/j.rse.2009.03.014
Class probabilities
First column is the site id, then one column of probability per class.
CI = 1 − (P1 − P2), where P1 and P2 are the two highest class probabilities. 0 means one class clearly wins; 1 means the top two are tied.
We build bootstrapped maps with prediction intervals, confusion indices and independent validation, so users know where a map can be trusted and where it can't.
Get in touch →Lin, L.I. (1989). Biometrics 45, 255–268. Malone, B.P., McBratney, A.B., Minasny, B. (2011). Geoderma 160, 614–626. Burrough, P.A., van Gaans, P.F.M., Hootsmans, R. (1997). Geoderma 77, 115–135. Open R implementation: rafikisol package.
Soil temperature waves
Daily and seasonal temperature swings at the surface fade and arrive later the deeper you go. See how soil moisture, density and clay control how far heat travels.
Temperature through the day
Thermal diffusivity vs water content
Diffusivity peaks at intermediate moisture: water first bridges grains and conducts heat well, then extra water mostly adds heat capacity.
We model coupled heat and water movement through layered profiles and across hillslopes, from surface conditions down to depth.
Get in touch →Uniform soil with sinusoidal surface temperature (Carslaw & Jaeger). Thermal conductivity: Campbell, G.S. (1985). Soil Physics with BASIC. Heat capacity: de Vries (1963).
Surface energy balance
Follow the sun's energy at a bare soil surface: how much heats the air, how much evaporates water, and how much flows into the ground. Wet the soil and watch evaporation take over.
Where net radiation goes
Partitioning as the soil wets up
We map energy and water budgets across fields and catchments, linking soil properties, moisture and terrain to where energy is stored and dissipated.
Get in touch →Steady-state bare-soil balance Rn = H + LE + G solved for surface temperature; neutral aerodynamic resistance at 2 m (z₀ = 1 cm); evaporation by the moisture-availability (β = θ/θfc) method; G by conduction over the top 10 cm (Campbell conductivity); sky emissivity after Brutsaert (1975). A teaching model, not a field estimate.
Evapotranspiration and crop water use
Daily reference evapotranspiration (ET₀) by the FAO-56 Penman–Monteith method, with Hargreaves for comparison, then crop water use and how long the soil's water lasts before the crop needs irrigating. Every step of the calculation is shown.
1. Place and day
2. Weather data you have
Use values for one day (or averages for a period). Missing data are filled in the FAO-56 way: dew point taken as the minimum temperature, wind as 2 m/s, and radiation from the daily temperature range. Each substitution adds uncertainty, and the results say which were used.
Reference evapotranspiration
3. Crop water use and irrigation interval
Typical available water: sands 50–100, loams 120–200, clays 150–200 mm per metre. The rooting depth starts at the FAO-56 typical value for the crop; shallower soils, compaction or young plants make it less.
We run daily soil water balances with measured or gridded weather, soil maps and crop calendars, for irrigation planning and water accounting.
Get in touch →Allen, R.G., Pereira, L.S., Raes, D., Smith, M. (1998). Crop Evapotranspiration: Guidelines for Computing Crop Water Requirements. FAO Irrigation and Drainage Paper 56. FAO, Rome. Equations 7–53 (ET₀), Table 12 (Kc; the midpoint where a range is given, and the crop-group initial value where no crop value is listed), Table 22 (rooting depth and depletion fraction), equations 82–84 (TAW, RAW). Checked against FAO-56 Example 18 (Brussels, 6 July: ET₀ 3.9 mm/day). Hargreaves, G.H., Samani, Z.A. (1985). Reference crop evapotranspiration from temperature. Applied Engineering in Agriculture 1, 96–99.
Frost depth
How deep the ground freezes in a winter, from the freezing index, the surface and the soil's density and moisture. Uses the two-phase Neumann solution, the exact basis of the modified Berggren equation, so the Stefan estimate and the Berggren λ correction are shown alongside.
1. Winter
2. Soil
The n-factor converts the air freezing index to the surface (values from TM 5-852-6, Table 2-4). Snow cover insulates the ground and can make frost much shallower; this calculation assumes the surface is kept clear or has no lasting snow. Water content is the average over the freezing depth in winter.
Frost depth
We model freezing and thawing through layered soils with snow, moisture and heat flow, for foundations, pipelines and slopes.
Get in touch →Method: two-phase Neumann solution for freezing of a semi-infinite medium (Carslaw, H.S., Jaeger, J.C. (1959). Conduction of Heat in Solids, 2nd ed. Oxford University Press), which underlies the modified Berggren equation (Aldrich, H.P., Paynter, H.M. (1953). Analytical studies of freezing and thawing of soils. Arctic Construction and Frost Effects Laboratory, Technical Report 42). Thermal conductivity: Kersten, M.S. (1949). Laboratory Research for the Determination of the Thermal Properties of Soils. University of Minnesota Engineering Experiment Station; reproduces the published example values for silt (0.72 and 0.80 BTU/h·ft·°F unfrozen and frozen) and is uncertain by about ±25%. Heat capacity, latent heat and n-factors: US Army (1988). Arctic and Subarctic Construction: Calculation Methods for Determination of Depths of Freeze and Thaw in Soils, TM 5-852-6.
Day length and solar radiation
Day length, sunrise and sunset, the noon sun angle and the solar energy reaching the top of the atmosphere and the ground on a clear day, for any latitude and date, with the whole year shown.
Sunrise and sunset are in local solar time (solar noon = 12:00); clock time differs with your position in the time zone and daylight saving. Day length is from sunrise to sunset of the sun's centre, without refraction, as in FAO-56, so published sunrise tables give a few minutes more.
Day length through the year
Daily solar radiation
Allen, R.G., Pereira, L.S., Raes, D., Smith, M. (1998). FAO Irrigation and Drainage Paper 56, equations 21–25, 34 and 37. Checked against FAO-56 Example 8 (20° S, 3 September: Ra 32.2 MJ/m² per day).
Pedodiversity
Measure the taxonomic diversity of soils in an area: how many soil classes there are, and how evenly the area is shared between them. Compare several sites at once.
Your soil classes
A column named class is required (any classification: soil orders, reference soil groups, series, map units). Add area or count to weight each row; without one, every row counts once (for example, one row per profile). Add site to compare several areas.
Indices by site
Indices
Share of each class
Reading the indices
Richness counts classes. Shannon (H′) rises with both more classes and more even shares. Evenness (J′) is Shannon relative to its maximum for that many classes, from 0 (one class dominates) to 1 (all equal). Simpson is the chance that two random points fall in different classes.
Hill numbers give the "effective number of classes": N1 weights classes by their share, N2 counts only the common ones. Dominance is the share of the most common class.
Taxonomic indices treat every class as equally different and ignore where soils sit and how they change with depth. Our approach measures pedodiversity from continuous soil properties in three dimensions, as the organisation of energy across the soil system.
Get in touch →Ibáñez, J.J., De-Alba, S., Bermúdez, F.F., García-Álvarez, A. (1995). Pedodiversity: concepts and measures. Catena 24, 215–232. Hill, M.O. (1973). Ecology 54, 427–432. Thermodynamic approach: Flynn, T. (2026). Preprint. doi:10.32942/X2H66M.