Does Cloud Seeding Pay Off for Farmers? North Dakota’s 2019 Sums

On paper it pays handsomely: a 2019 North Dakota State University report put the direct gain to crop producers in seeded counties at $28.2 million to $48.8 million a year (Bangsund & Hodur, 2019). The programme cost about $909,000 a year over 2008–2017 (Bangsund & Hodur, 2019). Those sums are modelled, not measured: they assume seeding cut hail losses by 45% and raised growing-season rain by 5% or 10% (Bangsund & Hodur, 2019). Independent reviews say the effects of seeding, especially on hail, remain hard to establish (U.S. Government Accountability Office, 2024; World Meteorological Organization, 2025).

Below, we explain what cloud seeding is meant to do over North Dakota’s fields, how the report turned assumed effects into dollars and which of its figures need care. We then set those assumptions against independent studies, federal and international reviews and the programme’s smaller footprint today. Winter seeding for water supply elsewhere appears only as evidence on whether seeding works, and claims about weather control are not covered. The original study is summarised in Cabdirect archive record 20193399635.

What Is Cloud Seeding Meant to Do Over North Dakota’s Crops?

In North Dakota, cloud seeding is meant to reduce hail damage to crops and to add rain during the growing season (North Dakota Department of Water Resources, 2026). The state programme uses glaciogenic seeding of warm-season clouds, a method that relies on agents such as silver iodide or dry ice (U.S. Government Accountability Office, 2024). Some form of seeding has been carried out in the state since the early 1950s (Bangsund & Hodur, 2019).

Hail is the main reason for the programme: the 2019 report describes North Dakota as consistently among the top states for crop hail damage (Bangsund & Hodur, 2019), and federal crop insurance hail indemnities in the state reached $59.2 million in 2024 (USDA Risk Management Agency, 2026).

From 2008 to 2017 the North Dakota Cloud Modification Project covered Bowman, McKenzie, Mountrail, Ward and Williams counties, part of Slope County and, from 2015, Burke County (Bangsund & Hodur, 2019). Only 41% of Slope County’s cropland was seeded (Bangsund & Hodur, 2019). Programme costs averaged about $909,000 a year over that decade (Bangsund & Hodur, 2019).

Whether seeding really suppresses hail has long been debated. The report leans on later analyses of insurance data as evidence that seeding cut hail damage (Bangsund & Hodur, 2019; Smith et al., 1997), while the World Meteorological Organization (WMO) says trials on hailstorms are not yet enough to accept the technology (World Meteorological Organization, 2025).

How Did the 2019 Study Put a Dollar Value on Hail and Rain?

The study priced two assumed effects of seeding, fewer hail losses and more growing-season rain, for nine crops on 20.4 million acres a year, in two assessments: the counties actually seeded and an imagined statewide programme (Bangsund & Hodur, 2019).

For hail, the authors applied crop-hail loss-cost ratios from National Crop Insurance Services to crop values and multiplied the result by a hail-reduction factor, 0.45 for most crops, taken from a 1997 analysis that found target-area loss ratios about 45% below expectation (Bangsund & Hodur, 2019; Smith et al., 1997).

For rain, the study modelled a 5% and a 10% rise in growing-season rainfall, with yield responses from relationships published in 1974; in the 5% case wheat gained about 1.5 bushels an acre (Bangsund & Hodur, 2019).

Statewide results were cut for lower crop prices, and the direct benefits were run through the IMPLAN input-output model to estimate wider business activity (Bangsund & Hodur, 2019).

Table 1. Key assumptions behind the 2019 estimates and where they came from

Assumption	Value used	Basis stated in the report
Cut in hail losses, most crops	45%	Smith et al. (1997), an exploratory analysis of crop-hail insurance data (Bangsund & Hodur, 2019; Smith et al., 1997)
Cut in hail losses, soybeans and dry beans	30%	Derived from Sell & Leistritz (1998) (Bangsund & Hodur, 2019)
Cut in hail losses, alfalfa	22.5%	Authors’ assumption: half of the 45% factor (Bangsund & Hodur, 2019)
Extra growing-season rain, first scenario	5%	Unpublished resource board estimates from observer data, 1977–2018 (Bangsund & Hodur, 2019)
Extra growing-season rain, second scenario	10%	No empirical basis given (Bangsund & Hodur, 2019)
Yield response to extra rain	1974 relationships	Johnson et al. (1974), rescaled to 2013–2017 yields (Bangsund & Hodur, 2019)
Wider economic effects	Input-output multipliers	IMPLAN model, benefits entered as household income (Bangsund & Hodur, 2019)

What Did the Study Estimate?

For the seeded counties, the study estimated $6.9 million a year in avoided hail losses, or $3.00 per planted acre (Bangsund & Hodur, 2019). Added rain was worth $21.2 million a year in the 5% scenario and $41.9 million in the 10% scenario, mostly from wheat and canola (Bangsund & Hodur, 2019). Total direct benefits to producers came to $28.2 million or $48.8 million a year, equal to $12.20 or $21.16 per planted acre (Bangsund & Hodur, 2019).

Set against the $909,000 annual cost, those totals imply roughly 31 to 54 dollars of direct benefit per dollar spent, and gross business volume reached $55.5 million to $96.3 million a year (Bangsund & Hodur, 2019).

A breakeven test found that, in the seeded counties, benefits matched costs at between 2% and 13% of the expected effect, about $0.40 per planted acre (Bangsund & Hodur, 2019).

The 2019 estimates exceed those of every earlier study in the series, which the authors put down mainly to rising crop values, while the assumed extra rain was revised down (Bangsund & Hodur, 2019).

Table 2. Estimated annual benefits of cloud seeding for North Dakota crops, averages for 2008–2017

Measure	Seeded counties	Hypothetical statewide programme
Planted acres, nine crops	2.31 million	20.4 million
Avoided hail losses	$6.9 million ($3.00 per acre)	$117.5 million ($5.75 per acre)
Value of added rain, 5% / 10% scenario	$21.2m / $41.9m	$181.8m / $359.4m
Total direct benefit, 5% / 10%	$28.2m / $48.8m	$299.2m / $476.9m
Direct benefit per planted acre, 5% / 10%	$12.20 / $21.16	$14.65 / $23.35
Gross business volume, 5% / 10%	$55.5m / $96.3m	$590m / $940m
Annual programme cost	$909,000	About $4 million (board estimate)
Direct benefit per $1 of cost, 5% / 10%	About 31 / 54	About 75 / 119

How Much of the Result Rests on Assumptions?

Nearly all of it. The study did not measure hail or rain in seeded areas; it took effect sizes from earlier work and asked what they would be worth, while stating without a source that the science has shown strong causality (Bangsund & Hodur, 2019).

The hail factor is the weakest link: Smith and colleagues presented their 1997 work as exploratory (Smith et al., 1997), its lower loss ratios could not be linked directly to hail damage (Knowles & Skidmore, 2021), and the factors for soybeans, dry beans and alfalfa were never measured (Bangsund & Hodur, 2019).

The rain scenarios rest on unpublished estimates for the 5% case and on no stated evidence for the 10% case (Bangsund & Hodur, 2019). A later rain-gauge evaluation of the same programme, as summarised by the US Government Accountability Office (GAO), found an average increase of 3%, not conclusive in seven of nine cases (U.S. Government Accountability Office, 2024).

A benefit-cost ratio is only as reliable as the weather effect fed into it; here the dollars are precise and the effect is assumed.

Other choices cut both ways: the yield equations date from 1974, while counting property, rangeland and crops such as potatoes would raise the benefits; the board that runs the programme funded the work (Bangsund & Hodur, 2019).

The headline ratio also needs care: the report’s 119-to-1 return matches the 10% scenario, although the sentence before it refers to the more conservative case, which gives about 75 to 1 (Bangsund & Hodur, 2019).

What Does Independent Evidence Show?

Independent work suggests some benefit in North Dakota but gives no firm size for it. A 2021 panel study of county data for 1989–2018 found that participating counties harvested 3.87 more bushels of wheat per acre, about 13% above the mean (Knowles & Skidmore, 2021).

Counting wheat alone, the same study put the benefit-cost ratio at 36 to 37 to one, and noted that benefits go to farmers and insurers while costs fall on all taxpayers (Knowles & Skidmore, 2021).

Wider reviews are more cautious. The GAO found extra precipitation of 0 to 20% in the studies it reviewed but judged the return on investment unclear (U.S. Government Accountability Office, 2024), and the WMO finds an evidence-based causal link only for wintertime seeding over mountains (World Meteorological Organization, 2025).

Even the best winter experiments show modest effects: a randomised Wyoming project in 2008–2013 found no statistically significant effect, though modelling suggested about 1.5% more annual precipitation (Rasmussen et al., 2018), and an Idaho study quantified snowfall from seeding in three cases (Friedrich et al., 2020).

Is Silver Iodide Safe, and Who Oversees Seeding?

At the levels used so far, the available evidence suggests silver iodide poses no health or environmental concern (U.S. Government Accountability Office, 2024; World Meteorological Organization, 2025), although the GAO notes it is not known whether wider use would have effects (U.S. Government Accountability Office, 2024).

The WMO adds that larger quantities or new agents would need evaluation, and operational programmes run in more than 50 countries (World Meteorological Organization, 2025).

In the US, federal involvement is minimal and oversight rests mainly on reports to the National Oceanic and Atmospheric Administration (U.S. Government Accountability Office, 2024), filed before and after each weather modification activity (Code of Federal Regulations, 2026).

States set their own rules. North Dakota licenses weather modification (North Dakota Department of Water Resources, 2026), Tennessee banned cloud seeding in 2024 (U.S. Government Accountability Office, 2024), and Florida’s 2025 act on geoengineering and weather modification took effect on 1 July 2025 (The Florida Senate, 2025).

What Has Changed in North Dakota Since 2019?

The seeded area has shrunk. In November 2024 voters in Williams and Mountrail counties ended their weather modification authorities (KFYR-TV, 2024), and the state’s project page now lists only Bowman and part of Slope County (North Dakota Department of Water Resources, 2026).

State lawmakers have kept the programme going: a 2023 House bill on cloud seeding failed in the Senate (North Dakota Legislative Branch, 2023), and in February 2025 the Senate rejected a bill to repeal the weather modification chapter (North Dakota Legislative Branch, 2025).

The crops the programme protects have also changed: in 2026 soybeans covered 6.8 million planted acres in North Dakota against 5.5 million for all wheat (USDA National Agricultural Statistics Service, 2026), whereas wheat averaged 7.6 million acres in the study period (Bangsund & Hodur, 2019). Such shifts would change any updated estimate.

Hail remains costly: in 2024 federal hail indemnities in the state were largest for soybeans and wheat (USDA Risk Management Agency, 2026).

Frequently asked questions

How does cloud seeding work in North Dakota?

North Dakota’s programme seeds warm-season clouds with agents such as silver iodide or dry ice to cut hail damage and add rain. In 2019 it covered seven western counties, one only in part, but the state’s project page now lists just Bowman and part of Slope County (Bangsund & Hodur, 2019; North Dakota Department of Water Resources, 2026).

Is cloud seeding worth the money for farmers?

It depends on uncertain effects. The 2019 report put direct benefits at $28.2 million to $48.8 million a year against about $909,000 in costs, but assumed the seeding effects (Bangsund & Hodur, 2019). A 2021 wheat study found 36 to 37 to one; the GAO calls the return unclear (U.S. Government Accountability Office, 2024; Knowles & Skidmore, 2021).

Does cloud seeding stop hail?

That has not been shown. The 2019 report’s 45% cut in hail losses came from an exploratory look at insurance data, not a controlled trial (Bangsund & Hodur, 2019; Smith et al., 1997). The WMO judges past hailstorm experiments insufficient to accept the concept of hail suppression (World Meteorological Organization, 2025).

Is silver iodide from cloud seeding harmful?

Not at the amounts used so far, on current evidence. The GAO and WMO report no significant health or environmental effects from past operations, but the effect of much wider use is unknown and the EPA lacks data for a formal risk determination (U.S. Government Accountability Office, 2024; World Meteorological Organization, 2025).

Is cloud seeding banned in any US states?

Yes. Tennessee banned it in 2024, and by December 2024 the GAO counted ten states that had banned or considered banning it (U.S. Government Accountability Office, 2024). Florida has regulated weather modification since July 2025 (The Florida Senate, 2025). North Dakota’s Senate rejected curbs on it in 2023 and 2025 (North Dakota Legislative Branch, 2025; North Dakota Legislative Branch, 2023).

Sources

Bangsund, D. A., & Hodur, N. M. (2019). Economic impacts of cloud seeding on agricultural crops in North Dakota (Agribusiness & Applied Economics Report No. 791). Department of Agribusiness and Applied Economics, North Dakota State University. – Cabdirect archive record 20193399635
U.S. Government Accountability Office. (2024). Cloud seeding technology: Assessing effectiveness and other challenges (GAO-25-107328).
Knowles, S., & Skidmore, M. (2021). Cloud seeding and crops yields: Evaluation of the North Dakota Cloud Modification Project. Weather, Climate, and Society, 13(4), 885–898.
Smith, P. L., Johnson, L. R., Priegnitz, D. L., Boe, B. A., & Mielke, P. W., Jr. (1997). An exploratory analysis of crop hail insurance data for evidence of cloud seeding effects in North Dakota. Journal of Applied Meteorology, 36(5), 463–473.
World Meteorological Organization. (2025). WMO statement on weather modification.
Friedrich, K., Ikeda, K., Tessendorf, S. A., French, J. R., Rauber, R. M., Geerts, B., Xue, L., Rasmussen, R. M., Blestrud, D. R., Kunkel, M. L., Dawson, N., & Parkinson, S. (2020). Quantifying snowfall from orographic cloud seeding. Proceedings of the National Academy of Sciences, 117(10), 5190–5195.
Rasmussen, R. M., Tessendorf, S. A., Xue, L., Weeks, C., Ikeda, K., Landolt, S., Breed, D., Deshler, T., & Lawrence, B. (2018). Evaluation of the Wyoming Weather Modification Pilot Project (WWMPP) using two approaches: Traditional statistics and ensemble modeling. Journal of Applied Meteorology and Climatology, 57(11), 2639–2660.
North Dakota Department of Water Resources. (2026). North Dakota Cloud Modification Project.
KFYR-TV. (2024, November 6). No more cloud seeding in Williams and Mountrail Counties.
North Dakota Legislative Branch. (2025). Senate Bill 2106, 69th Legislative Assembly: Bill overview.
North Dakota Legislative Branch. (2023). House Bill 1166, 68th Legislative Assembly: Bill overview.
The Florida Senate. (2025). CS/CS/SB 56: Geoengineering and weather modification activities.
Code of Federal Regulations. (2026). Title 15, Part 908: Maintaining records and submitting reports on weather modification activities.
USDA National Agricultural Statistics Service. (2026). Acreage (June 2026).
USDA Risk Management Agency. (2026). Summary of Business: Cause of loss data files, 2024 and 2025 [Data set].

Daniel Brooks — author

Daniel Brooks is a contributor to Cabdirect’s Environment & Climate content. He holds an MSc in Climate Science and a BSc in Environmental Geography and is affiliated with Northfield Climate Evidence Lab, a research and communication organisation. His backgrou...