Case Study
This page is a working case study of the SoilScope methodology — our general method for measuring the yearly value living land produces — run at full depth over one real landscape: 58,768 acres of farmland and open space protected by the Marin Agricultural Land Trust (MALT) in Marin County, California, as listed in the state's conservation easement records. Inputs from satellites, the federal soil survey, rain maps and published studies are worked out cell by cell over 57,840 map cells of 64 by 64 metres, for each month of 2025. All 9 services carry a price from a public source. The state's records list 90 easement names; MALT itself reports 98, because one record, “Marin Agricultural Land Trust CE”, is a catch-all for several properties rather than one ranch. The newest easement, Tunnel Hill Ranch (2026), is not yet included.
Measurements first: what the land does, then what it is worth — $58.9M to $90.5M a year, a range across two choices still open — and how each number is worked out.
The method. We follow InVEST, from the Natural Capital Alliance (formerly the Natural Capital Project at Stanford). It works out what the land does in physical amounts and has no built-in prices, so every dollar comes from a public price. Where no local price exists, the best general public value is used.
Method v4.5 · satellite greenness: each month of 2025 · every price traced to a published source · every service → · why this number changed → · what we don’t know yet → · sources →
↑ New here? Start with the general methodology — how we measure the value of land, in plain EnglishWhat the land does comes before what it is worth. Each amount is labelled. Measured means a satellite or survey reading. Estimated from measurements means worked out from those readings, with published methods and a few working rules of our own, each named under Sources. The soil, flood, clean-water, wildlife-habitat, nitrogen, cooling and carbon dollars further down are these amounts times a price: the calculation refuses to run unless the two agree in every one of the 57,840 map cells.
58,768 acres under easement (the easements' deeded acreage). The map covers 23,691 hectares (58,542 acres) in 57,840 squares of land.
From the national land-cover map, read from satellite images: shrub and scrub 11,465 hectares, grassland and pasture 6,564 hectares, forest 4,900 hectares, wetland 356 hectares, developed land 352 hectares, and open water, bare ground and crops 54 hectares.
The national map's shrub class likely includes much grassland; Marin's finer vegetation map will sort this out in a later version.
Measured by the Landsat satellites on summer mornings (about 10:35): 57.8% of the land is cooler than comparable built-up land nearby. Forest is 7.22 °C cooler on average. Grassland is no cooler than the built-up land at that hour: it is 1.41 °C warmer on average, and 69.7% of grassland measures hotter, the dry-grass effect.
1,021 mm of rain in an average year (the 1991–2020 average) and 804 mm in 2025, averaged over the land, from Oregon State University's rain maps.
About 1.81 million short tons of soil a year stay on the land compared with heavier grazing on the steepest hills, or 3.00 million short tons compared with farming all of it. Worked out with the US Department of Agriculture's revised soil-loss equation from each square's rain, soil, slope and measured greenness.
In the standard 2-year storm (92 mm in a day) the land holds back about 5.00 million cubic metres more water than row crops would, or 5.17 million cubic metres if the steepest hills are compared with hard-grazed pasture instead.
About 73,552 acre-feet (90.7 million cubic metres) in a typical year, and 45,919 acre-feet on 2025's rain: rain minus what plants and soil give back to the air. It feeds the creeks, wells, Tomales Bay and the Nicasio, Soulajule and Stafford reservoirs.
About 54,600 kilograms less nitrogen a year reaches West Marin creeks and bays than if every slope were farmed, or 28,100 kilograms compared with heavier grazing on the steepest hills. Worked out from the California Water Board's nitrogen export rates for each kind of land and soil.
About 12,702 tons of carbon dioxide a year, held in the trees on 4,900 hectares of forest. Only forest keeps a carbon credit: field measurements on grazed grassland (the Nicasio plots in Marin and Point Blue's statewide re-sampling) show no net yearly gain.
These sentences are written from the data, not by hand. Where a measurement cuts against us, as with dry summer grassland or carbon outside forest, the words say so, and they would change if the readings changed.
MALT’s land is worth $58.9M to $90.5M a year ($58,900,948.56 to $90,526,444.81), counting all nine services. It is a range because two choices are still open, so we show all four combinations instead of picking one.
Pacific-region price. $9.18 a short ton: the figure for California, Oregon and Washington together in the US Department of Agriculture's Economic Research Service report, the level its authors call credible.
Marin County price. $15.95 a short ton: the same report's figure for Marin County alone, which its authors call credible only when added up over several states.
Heavier grazing: the land is compared with hard-grazed grass (cover factor 0.20) on the steepest hills, and with farmland everywhere else.
Farmland, as today's rule assumes: every slope is compared with row crops (cover factor 0.343).
| Soil price → Steepest hills ↓ |
Pacific-region price$9.18 a short ton | Marin County price$15.95 a short ton |
|---|---|---|
| Heavier grazing | Pacific-region price, heavier grazinglow end of the range$58.9M$58,900,948.56 a year · $1,002 an acreLikely range $35.7M–$128.7M | Marin County price, heavier grazing$71.1M$71,144,922.20 a year · $1,211 an acreLikely range $46.9M–$142.0M |
| Farmland (today's rule) | Pacific-region price, farmland$70.2M$70,208,978.21 a year · $1,195 an acreLikely range $38.0M–$148.1M | Marin County price, farmlandhigh end of the range$90.5M$90,526,444.81 a year · $1,540 an acreLikely range $50.7M–$175.5M |
Each combination’s likely range. We re-ran the whole method for each combination with every other price and measurement at its low end together, then at its high end together, keeping that combination’s soil price: the cover of the farmed land the soil is compared with (0.228 to 0.450 instead of 0.343); the flood price ($0.89 to $4.32 per cubic metre a year instead of $1.73) and the design storm (82 to 104 mm instead of 92 mm); the carbon price (the auction floor, $27.94, to $32.48 a ton); clean water (only the water that drains to the supply reservoirs, up to all of it at the water index’s average since 2018, $555.01 an acre-foot); wildlife habitat (the lower, then the higher, of the two global value tables); pollination ($52,384 to $157,838 a year); the nitrogen price ($9.63 to $38.50 a kilogram); and cooling (from nothing, where air conditioning is rare, up to the value of city green space). The low end of the flood price assumes a 0% discount rate, which has no source (the federal guidance, Circular A-4, asks for 3% and 7%); it is kept only as a cautious floor. Both ends are real re-runs of the method, not a statistical formula. Each service’s own range is under Every Service.
If the soil price is also moved across its whole sourced band ($9.18 to $16.83 a short ton), the likely ranges widen to $35.7M–$143.7M with heavier grazing on the steepest hills, and $38.0M–$179.0M with farmland.
We follow InVEST, from the Natural Capital Alliance (formerly the Natural Capital Project at Stanford). Its models work out what the land does in physical amounts: tons of soil, cubic metres of water, kilograms of nitrogen, degrees of cooling. They have no built-in prices, so each dollar here comes from a public price, listed with its source under Sources. Where no Marin price exists, the best general public value is used instead of leaving the service at nothing. Dollars below are lowest to highest across the four combinations.
Soil the cover keeps on the hills compared with farming, priced at the US Department of Agriculture's value of a ton of soil kept out of the water ($9.18 or $15.95 a short ton): $16.62M to $47.89M a year.
How: for each cell and month, the US Department of Agriculture’s revised soil-loss equation (the force of the rain, how easily the soil washes away, the slope, and the gap between farmland’s plant cover and the cell’s own) gives the short tons an acre the cover keeps in place, times 2.4711 acres a hectare and the price. Under the heavier-grazing choice the steepest hills are compared with hard-grazed grass instead of farmland. No credit on open water or on ground half or more sealed.
Likely range: the farmed cover 0.228 to 0.450; the soil price itself is one of the two open choices.
The water the land yields in a typical year, priced at California's raw-water market price, $336.14 an acre-foot (the Nasdaq Veles California Water Index, 12-month average, 2025 dollars): about $24.72M a year.
How: InVEST’s water-yield model. Rain in an average year (1991–2020) minus what plants and soil give back to the air, on the unpaved share of each cell, gives 73,552 acre-feet. The price is the mean of the Nasdaq Veles California Water Index’s 52 weekly prices from October 29, 2025 to October 21, 2026, with the 2026 weeks brought to 2025 dollars: $336.14 an acre-foot. It is a statewide raw-water price; Marin has no raw-water market of its own. Flood control prices holding back one storm; clean water prices the water of a whole year, so the two do not overlap.
Likely range: $9.80M, counting only the 29,155 acre-feet that drain to the supply reservoirs (Nicasio, Soulajule and Stafford Lake), to $40.82M, all of it at the index’s average since 2018, $555.01 an acre-foot.
Extra storm water held back in the standard 2-year storm compared with row crops, priced at the yearly cost of a detention basin that holds the same water ($1.73 per cubic metre): $8.66M to $8.95M a year.
How: the federal runoff method on the standard 2-year, 24-hour storm, 92 mm: the water this ground holds minus what row crops on the same soil would hold (poor pasture on the steepest hills under the heavier-grazing choice), times 10 cubic metres a hectare for each millimetre, times the yearly cost of the Environmental Protection Agency’s cheapest detention basin, spread over 50 years.
Likely range: the storm 82 to 104 mm and the price $0.89 to $4.32 per cubic metre a year.
Habitat by land cover at the Ecosystem Services Valuation Database's standard values (grassland, shrub and pasture $145.54, forest $401.79, wetland $7,764.62 a hectare a year): about $7.39M a year.
How: InVEST’s habitat model maps habitat but gives it no dollar value, so each hectare takes the Ecosystem Services Valuation Database’s value for its land cover (Brander and others, 2024, Table 1, 2020 values brought to 2025 dollars). Only the database’s two habitat services are used, never a whole-landscape total, so nothing is counted twice. Shrub takes the grassland value: the database has no shrub value, and much of what the national map calls shrub on these ranches is grassland.
| Shrub and scrub | 11,465 ha | $145.54 |
| Grassland and pasture | 6,564 ha | $145.54 |
| Forest | 4,900 ha | $401.79 |
| Wetland | 356 ha | $7,764.62 |
| Open water | 48 ha | $784.92 |
| Crops | 5 ha | $1.87 |
| Developed and bare ground | 353 ha | $0.00 |
Likely range: $5.95M to $43.34M, taking for each land cover the lower, then the higher, of this table and its 2012 edition (de Groot and others). The 2012 grassland figure rests on two studies, so the high end is shown here only.
Nitrogen kept out of the creeks compared with farmland, priced at $24.07 a kilogram, the value of the harm nitrogen does to creek life in a US national study (Sobota and others, 2015), taken from Europe's nitrogen assessment: $677,000 to $1.32M a year.
How: InVEST’s nutrient model counts the service as the difference in what washes out between the land as it is and the land changed. For each cell: the nitrogen farmland would send downstream minus what this land sends, by land cover and soil group, from the Central Coast Water Board’s Salinas River nitrogen study (Appendix F, Table 3). That is 28,100–54,600 kilograms a year. Under the heavier-grazing choice the steepest hills are compared with grazing land instead of farmland. Phosphorus is not counted on its own: it travels mostly on soil, which the soil price already covers.
Likely range: the price $9.63 to $38.50 a kilogram: $526,000 to $2.10M when every slope is compared with farmland.
Carbon dioxide taken up by forest each year (2.99 tons a hectare from California's forest inventory), at the California carbon auction price of $32.48 a ton. Other land: no yearly gain measured, so no carbon credit: about $413,000 a year.
How: forest cells only: 2.99 tons of carbon dioxide a hectare a year, the measured gain in live trees on California’s private forest land, adjusted by satellite greenness (a working rule), times the August 2026 California carbon auction price as a reference. Grazed grassland and shrub get no carbon credit: field measurements at Nicasio in Marin and Point Blue’s statewide re-sampling show no steady yearly gain.
Likely range: the price $27.94 (the auction floor) to $32.48 a ton.
Wild bees and other insects living on this land help pollinate the squash, pumpkins, melons, berries and apples grown on West Marin farms; MALT is credited with half of the wild-insect share of that harvest: about $78,500 a year.
How: InVEST’s pollination model gives a score, not dollars, and needs a map of farm fields that does not exist for Marin. So the dollars come from county records: the value of Marin’s insect-pollinated fruit and vegetables (the 2024 crop report and the 2022 farm census), times each crop’s share of harvest owed to insects, times the share done by wild insects rather than honey bees (Calderone, 2012). MALT holds more than half of Marin’s farmland and is credited with half. The total is spread evenly over the land in natural cover, a working rule.
Likely range: $52,384 to $157,838 a year.
Air pollution taken up by plants, at the Forest Service's value for rural California, $1.62 a hectare a year: about $38,800 a year.
How: $1.62 a hectare a year on every cell, the US Forest Service’s value for rural California (Nowak and others, 2014, Table 3). The rate counts all land, sealed ground included, and the easements lie outside towns, so the rural value applies throughout.
Likely range: one published value; it does not move.
Trees, wetlands and green cover keep the ground cooler. Priced with InVEST's energy-savings formula: the air-conditioning electricity saved in homes and buildings within 450 m, counting only homes that have air conditioning. Few people live that close, so it is small: about $11,900 a year.
How: InVEST’s cooling model values cooling as the air-conditioning electricity saved in nearby buildings. We run its energy-savings formula on our own Landsat measurements. For each cell measured cooler than comparable built-up land: the degrees of cooling, times 0.1 degree of air cooling for each degree of ground cooling (our judgment, from the Environmental Protection Agency’s heat-island figures), times 0.74 kilowatt-hours a square metre a degree (Santamouris and others, 2018), times $0.3111 a kilowatt-hour (California homes, 2025), times the 51% of homes with central air conditioning in Pacific Gas and Electric’s area, times the built-up land within 450 metres, shared out by area. Land that measures hotter is not charged.
Likely range: nothing, because air conditioning is rare in the coastal fog belt and InVEST says not to use the formula there, up to about $116,000 at the value of city green space.
We are working with the tools and the data we have. Here, in plain words, is what the range above cannot see yet.
We have no ranch-by-ranch records of how the land is looked after: where cattle grazed and how hard, where compost was spread, what was planted or restored. The calculation sees the land from above; it cannot see the work. Practices such as compost can change the soil, but without records the calculation cannot credit them.
The calculation has no lab-tested soil samples from these ranches. Slope and soil group come from the federal soil survey’s maps, how easily the soil washes away from a formula of our own, and plant cover from satellites. Soil tested in a lab would check those readings on the ground.
What a ton of soil kept on the land is worth ($9.18 or $15.95 a short ton), and what the steepest hills, over 30% slope, would become without protection (heavier grazing or farmland). Together they make the range: see the four combinations.
Where no Marin price exists, the best general public value is used: wildlife habitat from a worldwide table by land cover, nitrogen from a US national study that takes its figures from Europe, clean water from California’s statewide water market. Pollination and cooling rest on small local numbers. Each one’s likely range is under Every Service.
Wildlife habitat is valued by what covers the land, not by how healthy that habitat is. The national map's shrub class likely includes much grassland; Marin's finer vegetation map will sort this out in a later version.
This page is about Marin. The click-anywhere world map prices the same nine services, but outside the United States the land cover is assumed until the world land-cover map is installed, and the prices are United States prices, so the dollars there are rough.
Every price has a published source. These steps inside the formulas do not; they are our own working rules, and each is named wherever it is used (the full list, under Sources):
The word methodology carries weight. It implies rigor. We use it deliberately, but also honestly: what you see here is scaffolding — a structured framework for measuring ecological value, not a finished building. These numbers are only as trustworthy as the data that feeds them, and right now, that data comes from satellites, public soil and rain maps and published studies, run through a calculation we built ourselves.
That is not enough. A single data stream can be noisy, biased, or simply wrong — clouds blind optical sensors, algorithms carry assumptions, and models are approximations of reality, never reality itself. We know this.
This is why we are building an oracle. Not a single pipeline, but 3–5 independent data streams that must all agree before a value is recorded as truth:
Drones went from military hardware to sub-$500 multispectral cameras in under a decade. Ground robotics is on the same curve — autonomous units that walk a ranch, collect samples, and return data without human intervention. As these tools get cheaper, the oracle gets denser, and the distance between measurement and ground truth shrinks toward zero.
Each data point must correlate with the others. When satellite greenness says the land is healthy, the soil sensors should confirm it. When the model says carbon is being sequestered, the lab assay should show rising organic matter. When all five streams converge on the same answer, we have something approaching truth. When they diverge, we have a signal to investigate — not to paper over.
The formulas below are the scaffolding. The oracle is what makes them honest. We publish this methodology not because it is finished, but because transparency is the only foundation worth building on.
Where the year’s value comes from: all nine services, each shown across the four combinations of the two open choices. The solid part of each bar is that service’s lowest value across the four combinations; the pale part is what its highest adds. Only soil kept in place, flood control and nutrient filtering move with the choices; the lowest values come from different combinations, so the solid parts need not add up to the low end of the range. Soil kept in place and carbon are worked out for each month of 2025 from Sentinel-2 satellite greenness and averaged; flood control is worked out once, on the standard 2-year, 24-hour storm; clean water on the rain of an average year.
How raw data becomes ecosystem value
Every figure we publish carries a method version; from v4.2 on, re-running that version reproduces it exactly, to the cent. Each change gets a new version and is listed here, with a before and after and the reason. From v4.4 the measurements come first and the dollars are shown as a range across the choices still open (see the four combinations). From v4.5 every service the InVEST method covers carries a public price (see Every Service), and every price traces to a published source, listed under Sources.
The figure shown over time: $155.1M → $165.5M → $97.3M → $80.1M → $25.9M–$56.9M → $58.9M–$90.5M
| Version | Date | Headline | What changed | Why |
|---|---|---|---|---|
| v4.0 | April 12, 2026 | $155.1M$155,064,085.25 | First valuation. 90 easements, 54,045 map cells of 64 m × 64 m, each valued from 12 monthly satellite sweeps of 2025 across nine services (soil kept in place, cooling, water supply, flood control, carbon, wildlife habitat, pollination, health, nutrient filtering). The largest pieces were soil kept in place ($83.5M) and cooling ($27.7M). | Starting point. |
| v4.1a | June 11, 2026 | $165.7M$165,726,328.30 | Fixed a counting bug. Ranches made of several parcels reused the same cell labels, so the export kept one cell per label and silently dropped the rest: about 3,795 real cells (McDowell Ranch went from 235 to 688 cells, Barboni Ranch from 1,271 to 2,444). Cells are now told apart by their map position. Cell count went from 54,045 to 57,840. | Real land was being left out. The total rose because missing acres came back, not because any price or formula changed. |
| v4.1b | June 11, 2026 | $165.5M$165,536,413.23 | Water supply and flood control had been worked out one month's rain at a time inside formulas meant for a whole year, then averaged. Both were recomputed from each cell's full year of rain. Water supply fell from $11.71M to $1.53M (in Marin, summer evaporation uses more water than falls as rain over the year). Flood control rose from $3.45M to $13.44M. The total moved by only 0.1%. | The split between services was wrong even though the total barely changed. |
| v4.2 | October 16, 2026 | $97.3M$97,290,390.91 | Two "compared with what?" corrections. Soil kept in place: the old method credited every hillside with stopping the erosion of bare, unplanted dirt. It now compares against what an easement actually prevents, conversion to vineyard, row crops or disturbed ground (a ground-cover factor of 0.35 instead of 1.0, where 1.0 means bare soil). Soil fell from $89.02M to $26.87M. Flood control: the storm-runoff formula is built for one storm, but it was being fed a whole year of rain. It now uses the standard design storm, the 2-year 24-hour storm (92 mm at the centre of the portfolio, from Atlas 14 of the National Oceanic and Atmospheric Administration, the United States weather agency), valued once a year at the cost of building the same amount of storage. Flood fell from $13.44M to $7.35M. | A careful outside reviewer would ask "when is this hillside ever bare?" and "why is a year of rain treated as one storm?". Both choices overstated the value, so both were corrected, using the comparison the owner chose from a written list of options on 2026-10-16. |
| v4.2 (wording) | October 20, 2026 | $97.3M$97,290,390.91 | Wording only, the number did not change. The landing-page sentence "Hundreds of millions in ecosystem value every year" was replaced with "About $97 million in ecosystem value every year". | That sentence was never a computed figure. It dates from when the computed total was $155–166 million, so it overstated even then, and from 2026-10-16 it sat beside a $97.3M headline. It was corrected four days later. |
| v4.3 | October 21, 2026 | $80.1M (likely range $69.1M–$92.0M)$80,105,679.29 | Cooling is now measured instead of assumed. Before, every unpaved spot was assumed to be 4 °C cooler than developed land. Now each cell's summer ground-surface temperature is measured by the Landsat 8 and Landsat 9 satellites (June to September, 2022–2025, 85 images on 64 dates) and compared with developed land within 15 km that sits the same distance from the ocean and at a similar elevation. The credit is capped at 4 °C cooler, because the dollars-per-degree rate is set so that 4 °C equals the $1,200 per hectare value used for forest. Unbuilt land that measures hotter than the developed land it is compared with earns $0, with no penalty. Cooling fell from $29.69M to $12.51M. No easement went up. Tree-covered land mostly measures more than 4 °C cooler than comparable developed land, so it kept most of its credit (forest cells $6.24M → $5.51M); open, dry grazing land, which measures about as hot as nearby towns on summer mornings, lost most of its credit (grassland cells $23.42M → $6.95M). The number now carries its method version and data window, re-running the method reproduces it to the cent, and it is published with a likely range. | The 4 °C figure was an assumption applied everywhere. The measurement showed tree cover really is much cooler than developed land, but dry summer grassland usually is not. Capping at 4 °C keeps the credit inside the forest value the rate is built on. |
| v4.4 | October 24, 2026 | $25.9M–$56.9M a year (range across the open choices)$25,892,455.39 – $56,879,579.84 | Every price and input was checked against its original source on October 22, 2026. Outright errors were fixed: soil was worked out in tons per acre but counted per hectare; one flat rainfall figure became each cell’s own; the 1978 slope curve became the 1997 one; the water figure’s rain came from a gauge in the wrong county; a file-reading bug garbled the satellite greenness for nearly a third of the easement records; and flood control now credits only the extra water the land holds compared with row crops. Prices that traced to no study were taken out of the dollars, and those services were shown as measurements instead. Health became cleaner air, and carbon is counted on forest only. | The $80.1M figure could be reproduced exactly but not defended. Fixing it showed that two judgment calls, both still open, move the figure by $31.0M a year: what a ton of soil kept on the land is worth and what the steepest hills would become without protection. So the measurements came first and the dollars were shown as a range across both choices: $25.9M to $56.9M. |
| v4.5 | October 24, 2026 | $58.9M–$90.5M a year (range across the open choices)$58,900,948.56 – $90,526,444.81 | Every service the Stanford method (InVEST) covers is now priced with a public value. The Los Angeles County stormwater fee is gone; it took $0.12M a year off the total. Clean water: California's raw-water market price, $336.14 an acre-foot in 2025 dollars, $24.7M a year. Wildlife habitat: the Ecosystem Services Valuation Database's 2024 values by land cover, $7.4M. Nutrient filtering: nitrogen kept out of the creeks compared with farmland, at $24.07 a kilogram, $1.32M at the internal default. Pollination ($0.08M) and cooling ($0.01M): Marin's crop values, and the air-conditioning energy saved nearby. Soil kept in place, flood control, forest carbon and cleaner air are unchanged ($47.89M, $8.66M, $0.41M and $0.04M at the internal default). | Owner decision on 2026-10-24 to price every service the Stanford (InVEST) method covers, using standard public values where no local price exists, and to drop the Los Angeles stormwater fee, which is not part of that method. InVEST, from the Natural Capital Alliance (formerly the Natural Capital Project at Stanford), works out amounts and has no built-in prices, so each dollar comes from a public price, listed with its source under Sources. The figure stays a range across the two open choices: the soil price (Marin County value $15.95 a short ton or Pacific-region value $9.18 a short ton) and whether the steepest hills would be farmed or grazed harder ($71.1M with heavier grazing at the Marin County price). History of the figure shown: $155.1M, $165.5M, $97.3M, $80.1M, $25.9M–$56.9M, $58.9M–$90.5M. |
Locking began with v4.3. v4.5 is checked three ways before it is published: run with every v4.3 setting, the same calculation reproduces $80,105,679.29 to the cent; run with the v4.4 settings, it reproduces $56,879,579.84 and the published v4.4 files byte for byte; and a second calculation, written separately in another programming language, reproduces all four combinations and every one of the 57,840 cells to the cent. The v4.2 figure has also been re-run from the stored measurements and reproduced to the cent; the earlier figures are taken from the saved data files of their time. Each entry is tied to a saved version of the engine and of this site.
Click any input to see what it captures, how it is used, and what it shows for MALT’s conservation easements in Marin County. Where a step rests on a working rule of our own rather than a published source, the card says so.
The sourced constants in the v4.5 calculation, each with its value, units, dollar year, source, table or page, link and the date we retrieved it. The working rules that have no outside source are listed after the table. Each raw download is kept with a digital fingerprint, so anyone can check that the file we used is the file the source publishes. The main values on this page are filled in from the same data file the calculation writes.
| What | Value | Source | Retrieved |
|---|---|---|---|
| Soil kept in place | |||
| Price of soil kept in place | $9.18 (Pacific region) or $15.95 (Marin County) a short ton (2,000 lb) a year, 2025 dollars, one of the two open choices; sourced band $9.18–$16.83 | Hansen and Ribaudo (2008), Economic Measures of Soil Conservation Benefits, US Department of Agriculture Economic Research Service, Technical Bulletin 1922, county table (appendix table 2): Marin County, $8.53 a ton in 2000 dollars, the sum of all 13 categories the report adds up. Low: the Pacific region (California, Oregon, Washington), $4.91, the level the authors call credible. High: the top of Marin’s $5–9 class in Figure 5, $9. Converted with the price index below. https://www.ers.usda.gov/media/8429/tb-1922.pdf · https://www.ers.usda.gov/media/14588/appendix-table-2.xls | 2026-10-23 |
| Price index | 2000→2025 ×1.8696; 2010→2025 ×1.4764; 1997→2025 ×2.0059 | US Bureau of Labor Statistics, consumer price index for all urban consumers, US city average, annual averages, series CUUR0000SA0 (2025 average 321.943, the mean of the 11 months published). https://api.bls.gov/publicAPI/v2/timeseries/data/ | 2026-10-23 |
| Rainfall-erosion factor | 57.42 to 145.62 across nine Marin rain bands (533–1,265 mm a year); US units, hundreds of foot-ton-force-inch per acre-hour-year | US Department of Agriculture, official erosion database for its revised soil-loss model, California weather file “CA climate120303”, the nine Marin County records. https://fargo.nserl.purdue.edu/RUSLE2_ftp/Climate_data/CA%20climate120303.zip | 2026-10-23 |
| Average-year rain | 1991–2020 normal, millimetres a year, for each cell, 800-metre grid | PRISM Climate Group, Oregon State University. https://data.prism.oregonstate.edu/normals/us/800m/ppt/monthly/prism_ppt_us_30s_2020_avg_30y.zip | 2026-10-23 |
| Slope steepness curve | S = 10.8 sin θ + 0.03 under 9%; 16.8 sin θ − 0.50 at 9% or more; support practice 1 | Renard and others (1997), Predicting Soil Erosion by Water, US Department of Agriculture Agriculture Handbook 703, p.107, equations 4-4 and 4-5. https://www.ars.usda.gov/ARSUserFiles/60600505/RUSLE/AH_703%20-%20Predicting%20Soil%20Erosion%20by%20Water%20-%20A%20Guide%20to%20Conservation%20Planning%20With%20the%20RUSLE%20(RUSLE).pdf | 2026-10-23 |
| Slope and soil group | representative slope (%) and hydrologic soil group of the main soil in each map unit | US Department of Agriculture soil survey, Marin County (survey version of September 16, 2026), through Soil Data Access. https://sdmdataaccess.sc.egov.usda.gov/ | April 2026 sweep; gaps re-queried 2026-10-23 |
| Cover of the farmed comparison | 0.343 (range 0.228–0.450), no units | Tetra Tech for the North Coast Regional Water Quality Control Board (2019), Laguna de Santa Rosa: Linkage Analysis for Sediment Impairments, Appendices, Table 1-2, p.1-8: annual cropland 0.343, vineyards and orchards 0.228. High: Wischmeier and Smith (1978), Predicting Rainfall Erosion Losses, Agriculture Handbook 537, Table 10, p.32 (no canopy, bare ground). https://www.waterboards.ca.gov/northcoast/water_issues/programs/tmdls/laguna_de_santa_rosa/pdf/251030/sed-budg-appx.pdf · https://www.ars.usda.gov/ARSUserFiles/60600505/RUSLE/AH_537%20Predicting%20Rainfall%20Soil%20Losses.pdf | 2026-10-22; 2026-10-23 |
| Units | 2.4710538 acres a hectare; a ton is 2,000 lb; 10 cubic metres a hectare per mm; 1,233.48 cubic metres an acre-foot; 107,639 square feet a hectare | National Institute of Standards and Technology, Handbook 44 (2012), Appendix C. https://nvlpubs.nist.gov/nistpubs/Legacy/hb/nisthandbook44-2012.pdf | 2026-10-23 |
| Flood control | |||
| Design storm | 92 mm in 24 hours, 2-year storm; range 82–104 mm (90% confidence) | National Weather Service, National Oceanic and Atmospheric Administration Atlas 14, Volume 6 (California), version 2, at 38.10° N, 122.75° W, partial-duration series. https://hdsc.nws.noaa.gov/cgi-bin/hdsc/new/cgi_readH5.py?lat=38.10&lon=-122.75&type=pf&data=depth&units=metric&series=pds | 2026-10-23 |
| Runoff method and curve numbers | grassland and pasture fair 49/69/79/84; grazed woods fair 36/60/73/79; brush fair 35/56/70/77; row crops, straight row, good 67/78/85/89 (soil groups A/B/C/D) | US Soil Conservation Service (1986), Urban Hydrology for Small Watersheds, Technical Release 55, 2nd edition, equations 2-1 to 2-4 and Tables 2-2a to 2-2c (copy of the federal document). https://ponce.sdsu.edu/tr55.pdf | 2026-10-23 |
| Two-grade soils | use the second (undrained) grade | US Department of Agriculture, National Engineering Handbook Part 630, Chapter 7 (2009), p.7-2. https://directives.nrcs.usda.gov/sites/default/files2/1712930597/11905.pdf | 2026-10-23 |
| Detention basin cost | $35.42 a cubic metre (2025 dollars) = $0.50 a cubic foot (1997 dollars); high end $1.00 a cubic foot | Environmental Protection Agency (1999), Preliminary Data Summary of Urban Storm Water Best Management Practices, report 821-R-99-012, Table 6-1, p.6-3. Design, contingency and land are not included, so this is a floor. https://www.epa.gov/sites/default/files/2015-11/documents/urban-stormwater-bmps_preliminary-study_1999.pdf | 2026-10-23 |
| Yearly flood price | $1.73 per cubic metre a year (range $0.89–$4.32), 2025 dollars | Basin cost × [0.03 ÷ (1 − 1.03−50) + 0.01]. The 3% discount rate: Office of Management and Budget, Circular A-4 (2003), reinstated by Executive Order 14192 (2025). Upkeep 1% a year: the same Environmental Protection Agency report, Table 6-10 (“<1%”). Range: low-end basin over 50 years with 0.5% upkeep and no discounting; high-end basin over 30 years. The 50-year and 30-year lives are our assumption; no published source found yet. The low end’s 0% discount rate has no source either (Circular A-4 asks for 3% and 7%); it is kept only as a cautious floor. https://obamawhitehouse.archives.gov/omb/circulars_a004_a-4/ · https://www.govinfo.gov/content/pkg/FR-2025-02-06/pdf/2025-02345.pdf | 2026-10-23 |
| Carbon held in trees | |||
| Carbon price | $32.48 a metric ton of carbon dioxide (low $27.94, the auction floor), August 2026 dollars | California Air Resources Board and Québec, August 2026 Joint Auction #48, Summary Results Report, Table 4 (settlement and reserve prices). A reference price only: ranch carbon cannot be sold into this market. https://ww2.arb.ca.gov/sites/default/files/2026-08/nc-aug_2026_summary_results_report.pdf | 2026-10-23 |
| Forest carbon rate | 2.99 tons of carbon dioxide a hectare a year (95% interval 2.50–3.48); every other land cover 0: no measured yearly gain | Christensen and others (2019), AB 1504 California Forest Ecosystem and Harvested Wood Product Carbon Inventory: 2017 Reporting Period, US Forest Service with the California Department of Forestry and Fire Protection and the California Board of Forestry, Table 4.4a, private non-corporate land, net change in live trees above ground. https://bof.fire.ca.gov/media/8026/4-final_1504_forest_ecosys_hwp_c_2017_13feb19_full.pdf | 2026-10-23 |
| Why other land covers get 0 | no measured yearly gain | Ryals and Silver (2013), Ecological Applications 23:46–59, Appendix A: unamended grassland plots at Nicasio, Marin, lose carbon. https://doi.org/10.6084/m9.figshare.3517898. Veloz and others (2022), Mapping California Rangeland Soil Carbon, Point Blue Conservation Science, p.17: average change −0.59 tons of carbon a hectare a year. https://www.pointblue.org/wp-content/uploads/2022/09/Statewide-SoilCarbon_TechnicalReport_8.30.22_FINAL-5.pdf. Ma and others (2007), Agricultural and Forest Meteorology 147:157–171. https://doi.org/10.1016/j.agrformet.2007.07.008 | 2026-10-22; 2026-10-23 |
| Cleaner air | |||
| Cleaner air | $1.62 a hectare of land a year (2025 dollars) = $1.1 (2010 dollars) × 1.4764, on every cell. The rate counts all land, sealed ground included, so it is not reduced for sealed ground; the easements lie outside census urban areas, so the rural rate applies throughout. | Nowak, Hirabayashi, Bodine and Greenfield (2014), “Tree and forest effects on air quality and human health in the United States”, Environmental Pollution 193:119–129, Table 3, California, rural land. https://www.itreetools.org/documents/628/nrs_2014_nowak_001.pdf | 2026-10-23 |
| Clean water | |||
| Water price | $336.14 an acre-foot = $0.2725 a cubic metre, 2025 dollars; high end $555.01 an acre-foot | Nasdaq Veles California Water Index, weekly values in dollars an acre-foot: the mean of the 52 weeks from October 29, 2025 to October 21, 2026, each 2026 week brought to 2025 dollars with that month’s consumer price index (September carried for October). The index reflects “the commodity value of water at the source”: a statewide raw-water price, not a Marin sale price. High end: all 417 weekly values since October 31, 2018, in 2025 dollars. https://indexes.nasdaqomx.com/Index/HistoryData | 2026-10-24 |
| Supply-reservoir catchments (low end) | the land above the dams of Nicasio Reservoir, Soulajule Reservoir, Stafford Lake and Kent Lake | US Geological Survey, Network Linked Data Index, basin above each dam’s stream reach (5329855, 5329843, 1669839, 5330431), dams matched through the National Inventory of Dams. For example https://api.water.usgs.gov/nldi/linked-data/comid/5329855/basin | 2026-10-24 |
| Clean-water yield | evaporation curve; plant-water coefficient 2.0 for forest, 0.5 for grassland; times the unsealed share | Zhang, Dawes and Walker (2001), Water Resources Research 37:701–708 https://doi.org/10.1029/2000WR900325; coefficients as printed in Bradford, Zhang and Hairsine (2001), Technical Report 01/8, p.3. https://fao.org/forestry-fao/42679-06c5b4f34ca7eb9155af198cbfdf42498.pdf | 2026-10-23 |
| Potential evaporation | 903 mm in 2025, one point at 38.05° N, 122.75° W | Open-Meteo historical weather (daily reference evaporation by the United Nations Food and Agriculture Organization method). https://archive-api.open-meteo.com/v1/archive | April 2026 sweep |
| Wildlife habitat | |||
| Habitat value by land cover | a hectare a year, 2025 dollars: grassland, shrub and pasture $145.54; forest $401.79; wetland $7,764.62; open water $784.92; crops $1.87; developed and bare ground $0 | Brander, de Groot and others (2024), “Economic values for ecosystem services: A global synthesis and way forward”, Ecosystem Services 66:101606 (the Ecosystem Services Valuation Database), Table 1, p.8: maintenance of life cycles and of genetic diversity, by biome, in 2020 dollars, brought to 2025 dollars with the consumer price index. Shrub takes the rangeland value (a working rule). https://eprints.gla.ac.uk/322904/1/322904.pdf | 2026-10-24 |
| Habitat range ends | high: grassland, shrub and pasture $1,885.00, forest $1,338.44; low: wetland $3,811.91, open water $0 | de Groot, Brander and others (2012), “Global estimates of the value of ecosystems and their services in monetary units”, Ecosystem Services 1:50–61, Table 2, habitat services, 2007 dollars brought to 2025. The grassland figure is the mean of two studies. https://www.cbd.int/financial/values/g-valueglobal.pdf | 2026-10-24 |
| Nutrient filtering | |||
| Nitrogen washed off each kind of land | kilograms of nitrogen a hectare a year, soil groups A/B/C/D: farmland 2 / 3.52 / 5.09 / 6.61; pasture 0.8 / 1.92 / 3.08 / 4.2; range and grazing land 0.97 / 2.04 / 3.13 / 4.2; forest 0.68 / 1.21 / 1.77 / 2.3; wetland 0 | California Regional Water Quality Control Board, Central Coast Region (2013), Total Maximum Daily Loads for Nitrogen Compounds and Orthophosphate for the Lower Salinas River and Reclamation Canal Basin, Appendix F, Table 3. Cited by the InVEST nutrient guide. Shrub uses the grazing-land row (a working rule). https://www.waterboards.ca.gov/centralcoast/water_issues/programs/tmdl/docs/salinas/nutrients/att2_appendix_f.pdf | 2026-10-24 |
| Nitrogen price | $24.07 a kilogram (low $9.63, high $38.50), 2025 dollars | Sobota, Compton, McCrackin and Singh (2015), “Cost of reactive nitrogen release from human activities to the environment in the United States”, Environmental Research Letters 10:025006, Table 1, increased eutrophication of fresh water, $6.44 / $16.10 / $25.75 a kilogram in 2008 dollars, brought to 2025. The row takes its figures from Europe’s nitrogen assessment (van Grinsven and others, 2013). https://iopscience.iop.org/article/10.1088/1748-9326/10/2/025006 · https://doi.org/10.1021/es303217e | 2026-10-24 |
| Pollination | |||
| MALT pollination total | $78,624 a year (low $52,384, high $157,838), 2025 dollars, spread evenly over the land in natural cover (a working rule) | Marin County Department of Agriculture, Weights and Measures, 2024 Crop and Livestock Report (fruit and vegetables $2,430,000; county farmland 156,458.5 acres) https://marincarbonproject.org/wp-content/uploads/2025/07/250626_2025-Crop-Report_FINAL_Web-as-spreads-tagged.pdf; US Department of Agriculture, 2022 Census of Agriculture, California county tables (Marin crop acres and sales) https://www.nass.usda.gov/Publications/AgCensus/2022/Full_Report/Volume_1,_Chapter_2_County_Level/California/cav1.pdf; Calderone (2012), PLoS ONE 7:e37235, Table S12 (shares due to insects and to wild insects); Klein and others (2007), Proceedings of the Royal Society B 274:303–313 (olives); Marin Agricultural Land Trust, news, October 11, 2021 (more than half of Marin’s farmland). https://journals.plos.org/plosone/article/file?type=supplementary&id=10.1371/journal.pone.0037235.s003 | 2026-10-24 |
| Cooling | |||
| The formula | energy saved = degrees of air cooling × kilowatt-hours a square metre a degree × electricity price, on built-up land within 450 m, reduced to the share of homes with air conditioning | InVEST User Guide (Natural Capital Alliance), Urban Cooling model, energy-savings valuation. The guide says that if air conditioning is rare “this valuation metric should not be used”, which sets the low end at nothing. https://storage.googleapis.com/releases.naturalcapitalproject.org/invest-userguide/latest/en/urban_cooling_model.html | 2026-10-24 |
| Energy per degree | 0.74 kilowatt-hours a square metre a degree a year | Santamouris and others (2018), Energy and Buildings 166:154–164. https://iris.poliba.it/bitstream/11589/122669/3/post-print.pdf | 2026-10-24 |
| Electricity price | $0.3111 a kilowatt-hour, 2025 | US Energy Information Administration, Electric Power Monthly, Table 5.6.B, California homes, January–August 2025. https://www.eia.gov/electricity/monthly/epm_table_grapher.php?t=epmt_5_6_b | 2026-10-24 |
| Homes with air conditioning | 51% with central air conditioning | California Energy Commission (2021), 2019 Residential Appliance Saturation Study, Table ES-1, Pacific Gas and Electric’s area. Likely high for the coast. https://www.energy.ca.gov/sites/default/files/2021-08/CEC-200-2021-005-ES.pdf | 2026-10-24 |
| Air cooling for each degree of ground cooling | 0.1 (a working rule) | Our judgment from the Environmental Protection Agency (2008), Reducing Urban Heat Islands: Compendium of Strategies, chapter 1, Table 1 (surface 10–15 °C against air −1 to 3 °C by day). https://www.epa.gov/sites/default/files/2017-05/documents/reducing_urban_heat_islands_ch_1.pdf | 2026-10-24 |
| Cooling high end | $1,951.36 a hectare of city green space a year, scaled by the built-up land nearby against a half-built-up city (a working rule) | Elmqvist and others (2015), Current Opinion in Environmental Sustainability 14:101–108, Table 2, energy savings and temperature regulation, 2013 dollars brought to 2025. https://research.fs.usda.gov/download/treesearch/49282.pdf | 2026-10-24 |
| The method, and dollar years | |||
| The method | InVEST: physical amounts, no built-in prices | InVEST User Guide, Natural Capital Alliance (formerly the Natural Capital Project at Stanford), 2026, version 3.20.2: Annual Water Yield, Habitat Quality, Nutrient Delivery Ratio, Crop Pollination and Urban Cooling models. https://storage.googleapis.com/releases.naturalcapitalproject.org/invest-userguide/latest/en/index.html | 2026-10-24 |
| Monthly price index (2026) | each 2026 water-index week × 321.943 ÷ that month’s index | US Bureau of Labor Statistics, consumer price index for all urban consumers, monthly, January–September 2026, series CUUR0000SA0. https://api.bls.gov/publicAPI/v2/timeseries/data/ | 2026-10-23 |
| Measured inputs | |||
| Satellite greenness | each month of 2025; per pixel, the highest greenness over clear passes; averaged inside each easement outline | Copernicus Sentinel-2 surface reflectance (level 2A), through the Copernicus Data Space Ecosystem processing service. https://dataspace.copernicus.eu/ | 2026-10-23 |
| Land cover and sealed ground | 30-metre classes and sealed share | National Land Cover Database 2021, Multi-Resolution Land Characteristics Consortium. https://www.mrlc.gov/ | April 2026 sweep |
| Summer ground temperature | June–September 2022–2025, 85 images; built-up comparison land, and built-up land within 450 m for cooling | Landsat 8 and 9, Collection 2, Level 2 surface temperature (US Geological Survey), read from Microsoft Planetary Computer’s public copy; built-up land from European Space Agency WorldCover 2021. https://planetarycomputer.microsoft.com/ | October 2026 |
These steps are our own, most of them kept from the earlier method. We list them so you can see exactly what rests on our judgment rather than on a published source.