Shrink-swell clay soil by county: Ohio, Illinois, Texas

What this dataset is

This is our own calculation. NRCS does not publish a "share of county with high shrink-swell soil" figure. We built one from public USDA NRCS soil survey data (SSURGO), pulled through the Soil Data Access service on October 6, 2026^[1]. If you cite these numbers, please cite them as a Top Concrete Workers calculation from USDA NRCS SSURGO data.

Shrink-swell is how much a soil expands when wet and shrinks when dry. NRCS measures it as linear extensibility (LEP): the change in length of a soil clod as it dries from moist to oven-dry, stated as a %^[2]. NRCS classes are^[2]:

LEP	Shrink-swell class
Less than 3.0%	Low
3.0% to 5.9%	Moderate
6.0% to 8.9%	High
9.0% or more	Very high

NRCS states that soils rated moderate to very high "can damage buildings, roads, and other structures"^[2]. In southeast Texas, the Foundation Performance Association says foundation movement damage is most often caused by moisture changes in expansive soils^[4].

The county table

"High or very high" is the share of rated soil area with LEP of 6% or more. Housing age columns come from the Census ACS 2024 5-year estimates^[3]. The median year built is interpolated within the Census decade bins.

County (main city)	Rated soil acres	High or very high (LEP 6%+)	Very high (LEP 9%+)	Moderate or higher (LEP 3%+)	Homes built before 1980	Median year built
Dallas, TX (Dallas)	398,049	70.3%	45.0%	87.8%	44.9%	1983
Tarrant, TX (Fort Worth)	417,164	59.5%	33.1%	79.0%	32.6%	1990
Bexar, TX (San Antonio)	716,266	47.4%	21.3%	86.0%	36.3%	1990
Travis, TX (Austin)	492,136	44.6%	33.6%	88.3%	24.0%	1999
Harris, TX (Houston)	765,231	42.2%	21.0%	63.2%	37.1%	1988
Sangamon, IL (Springfield)	539,896	39.7%	0.0%	97.9%	61.8%	1974
Cook, IL (Chicago)	412,687	36.7%	0.0%	91.2%	73.2%	1961
DuPage, IL (Naperville, part of Aurora)	205,282	32.6%	0.0%	97.2%	53.5%	1978
Peoria, IL (Peoria)	376,441	22.9%	0.0%	78.5%	70.4%	1966
Will, IL (Joliet, part of Naperville)	515,569	22.2%	0.0%	91.0%	37.0%	1992
Lucas, OH (Toledo)	160,061	20.8%	0.0%	63.3%	72.9%	1963
Hamilton, OH (Cincinnati)	136,019	14.1%	0.3%	86.1%	72.6%	1963
Winnebago, IL (Rockford)	301,105	8.3%	0.0%	85.8%	63.4%	1971
Montgomery, OH (Dayton)	240,550	7.2%	0.0%	94.3%	73.6%	1966
Kane, IL (Aurora)	314,875	6.3%	0.0%	97.7%	45.3%	1985
Summit, OH (Akron)	218,885	2.3%	0.0%	33.6%	68.2%	1966
Franklin, OH (Columbus)	280,613	1.3%	0.0%	95.5%	51.5%	1979
Cuyahoga, OH (Cleveland)	177,718	1.2%	0.0%	77.3%	80.4%	1957

Soil columns: our calculation from USDA NRCS SSURGO via Soil Data Access^[1], with NRCS class limits^[2]. Housing columns: ACS 2024 5-year, table B25034^[3]; margins of error not shown.

What the numbers say

• Texas metro counties stand apart. All five have 42% to 70% of rated soil in the high or very high class, and 21% to 45% in very high alone. This is the soil story behind slab foundation repair in Texas. See expansive clay soil.
• Chicago area counties sit in the middle. Cook (37%) and DuPage (33%) have real high shrink-swell shares but no very high soil in our pull. Kane is much lower at 6%.
• Ohio is mostly moderate. Columbus and Cleveland area counties are near 1% high. The Ohio foundation story is mostly frost, water and older housing, not swelling clay. Cuyahoga County has the oldest housing in this table, with 80% of homes built before 1980. See frost heave.
• "Moderate or higher" is high almost everywhere. Most counties are above 75% on that measure. Moderate soils can still move, which is one reason drainage advice applies in every state.

Method

1. For each county's soil survey area (for example TX113 for Dallas County), we took every map unit component.
2. For each component, we found the highest representative LEP value (the SSURGO field lep_r) among soil layers that start in the top 150 cm.
3. We weighted each component by its share of the map unit (comppct_r) times the map unit's acres (muacres).
4. Share = weighted acres with a maximum LEP of 6% or more, divided by all weighted acres with an LEP value.
5. Components with no LEP data, mainly urban land and water, are left out. So the share describes the soil-rated part of the county, not every parcel. In heavily built counties, a large area may be mapped as urban land.

The query is plain SQL sent to the public Soil Data Access Tabular REST service at https://sdmdataaccess.sc.egov.usda.gov/Tabular/post.rest, with JSON output^[1]. Here is the full query:

SELECT l.areasymbol,
 SUM(CASE WHEN h.maxlep >= 6 THEN mu.muacres * c.comppct_r / 100.0 ELSE 0 END) AS ac_high_vhigh,
 SUM(CASE WHEN h.maxlep >= 9 THEN mu.muacres * c.comppct_r / 100.0 ELSE 0 END) AS ac_vhigh,
 SUM(CASE WHEN h.maxlep >= 3 THEN mu.muacres * c.comppct_r / 100.0 ELSE 0 END) AS ac_mod_plus,
 SUM(mu.muacres * c.comppct_r / 100.0) AS ac_rated
FROM legend l
 JOIN mapunit mu ON mu.lkey = l.lkey
 JOIN component c ON c.mukey = mu.mukey
 JOIN (SELECT cokey, MAX(lep_r) AS maxlep FROM chorizon
       WHERE hzdept_r < 150 AND lep_r IS NOT NULL GROUP BY cokey) h
   ON h.cokey = c.cokey
WHERE l.areasymbol IN ('OH049','OH035','OH061','OH113','OH095','OH153',
 'IL031','IL043','IL197','IL089','IL201','IL143','IL167',
 'TX201','TX113','TX439','TX029','TX453')
GROUP BY l.areasymbol

Each column is then divided by ac_rated to get the shares in the table.

Limits of this data

• It is a county average, not a test of your lot. Soil changes over short distances, and builders often bring in fill.
• We used the worst layer in the top 150 cm. A component counts as high if any layer starting in that depth is high. That leans toward higher shares than an average across layers would.
• LEP is a lab property. It says how much a soil can move, not how much it will. Actual movement depends on moisture swings, drainage, trees and how the foundation was designed.
• A high share does not predict damage to any one house, and a low share does not rule it out. Settlement, poor fill, leaks and frost cause damage on low shrink-swell soils too.
• Survey areas are close to, but not always identical to, county lines.

When to call a pro

• If your county ranks high and you see new cracks or sticking doors, a structural engineer can check whether the slab has moved.
• Before buying a house in a high shrink-swell county, ask for any past foundation repair records and an engineer's inspection.
• For new construction or an addition, a geotechnical engineer can test the soil on your actual lot.