Previous AZCIR reporting has revealed regulatory weaknesses in how Arizona manages its groundwater, including concerns from local residents about which individuals and groups are disproportionately depleting local aquifers. In the Hualapai Valley north of Kingman, we heard complaints about foreign-owned farms extracting substantial water—and profits—from a basin under threat.
To better understand the scope and scale of foreign ownership, and the role these entities have played in the area’s recent surge of industrial-scale agriculture, AZCIR spent several months analyzing various public records and datasets. Our investigation traced strings of LLCs and private investment firms operating in the northwestern part of the state.
As lawmakers debate how to incorporate local control into new groundwater regulations, this analysis aims to help inform that conversation. Here’s how we did it.
What data was used?
How did the data intersect with the groundwater basin?
How did you determine land use within the basin?
How did you prove a company was from out of state?
How can you show the movement of companies into the basin?
What data was used?
In Arizona, groundwater is considered a public good and cannot be owned by any individual. Across most of the state, it is governed only by the doctrine of “reasonable use,” meaning it may be drawn in any amount provided the water is being used productively. As a result, much of Arizona’s groundwater remains effectively unregulated.
The closest the state gets to establishing groundwater ownership in regulated areas is via usufruct rights—rights to use, rather than possess, the water. These rights ensure that those who were accessing groundwater in a particular area before regulations took effect can continue to do so, based on historical usage and where the water is applied.
The datasets used in this analysis were selected because they are representative of the two key parts of the issue: the water itself and the land above it.
Dataset 1: “The Water” – ADWR Well55 data
This Arizona Department of Water Resources (ADWR) dataset includes all well registrations in the state. It has information such as well owners, approximate locations, size, depth, use and construction details. Each entry also includes a reference number tied to a repository of publicly accessible documents held by ADWR. The repository contains information covering the various stages of a well’s history, from drilling logs and ownership transfers to inspections and court filings. Together, the Well55 data and document repository allowed AZCIR to verify ownership details for individual wells. This helped reporters identify the major water users in the basin, determine when they established operations or started drilling, and see how deep they drilled to reach the aquifer below.
Dataset 2: “The Land” – Mohave County Parcel Data
AZCIR gathered data on individual land parcels through a public records request to the Mohave County Assessor’s Office. The dataset includes parcel owners, registered addresses, parcel size in acres and parcel shapes for digital mapping purposes. It also contains land-use codes from the Arizona Department of Revenue (AZDOR). These codes are based on an on-site visit or GIS-aided survey to determine each property’s primary use, which is then tied to its appraised value, according to the AZDOR Property Use Code Manual.
How did the data intersect with the groundwater basin?
Because AZCIR’s parcel-level data included exact locations and shapes, reporters were able to overlay the Hualapai Valley Basin boundary to identify land that either fell within or intersected it. Using geographic information system software, reporters then isolated all parcels within the basin for further analysis.
We opted to include parcels that straddled the basin boundary in our dataset, meaning the total acreage reflected in our calculations is slightly larger than the basin’s actual area. Put simply, if any portion of a parcel lay within the basin, the full parcel acreage was counted. However, this did not affect the final analysis, which focuses on the owned shares of currently cultivated land. None of the parcels that were actively used for agriculture straddled the basin line.
How did you determine land use within the basin?
AZDOR land-use codes proved especially useful for verifying on-the-ground reports of nut tree orchards. Satellite-based tools, such as the USDA’s National Agricultural Statistics Service (NASS) Cropland Data Layer (CDL), were less effective for this purpose, because young nut trees can evade this type of satellite analysis until they mature. By merging parcel data with ADOR land-use data, AZCIR found that about 40% of the cropland in the Hualapai Basin is immature nut trees.
To isolate any cultivated cropland and exclude ranchland, which makes up a large portion of the basin, AZCIR focused on the following AZDOR codes:
- 4110, 4117 (Field Crops)
- 4210, 4211, 4215 (Mature Vineyards)
- 4360, 4361, 4362, 4363, 4367 (Immature Nut Trees)
- 4510, 4512 (High Density Agriculture)
Other agriculture codes did not show up in the basin data, indicating a lack of any substantial activity in those categories. One ranchland code did note some agricultural usage, but it was excluded from the final analysis because agriculture was not the primary purpose. This exclusion did not have a statistically significant effect on final calculations.
AZCIR relied on these granular codes rather than broader categories because much of the cultivated farmland in the basin was tagged as “commercial” agriculture, indicating that it’s intended for export, according to the Mohave County Assessor’s Office.
It is important to note that these codes denote only the primary use of the land. Portions of a parcel may contain sections of uncultivated land, while some cultivated areas may fall within parcels used primarily for other purposes. Smaller vineyards in the basin, for example, may be particularly underrepresented in the dataset.
Use of the DOR codes is meant to provide a broader view of parcels used primarily as cropland over multiple years. Due to business and environmental factors, the number of currently cultivated acres fluctuate year to year, particularly for operations with field crops.
To date, ADWR has not published precise, mappable data showing allowable cultivation within the Irrigation Non-Expansion Area (INA). Grandfathered rights applications are still being adjudicated.
How did you prove a company was from out of state?
Reporters reviewed company filings through the Arizona Corporation Commission (AZCC) and equivalent state agencies or company registrars in California, Nevada and other relevant states. From there, AZCIR established criteria to determine when a company could be considered an out-of-state interest. A company met that definition if either of the following applied:
- The entity is registered with the Arizona Corporation Commission (AZCC) as a Foreign LLC, a company formed in another state that has a branch—essentially, a subsidiary—in Arizona.
- Its registration address with the AZCC is outside of Arizona, supported by other evidence from further backgrounding.
The registration address, while typically indicative of where an LLC operates, can sometimes point instead to the agent used to file the company’s paperwork. Because these agents are frequently unaffiliated with the businesses themselves, at least one of the following also needed to be true in such cases:
- OpenCorporates (or a similar aggregator) shows the company is a direct subsidiary of an entity based outside Arizona.
- OpenCorporates (or a similar aggregator) shows the company is owned and managed by entities based outside Arizona.
- Additional research shows that the company advertises or identifies itself as being primarily based in another state.
- Additional research shows that the company focuses on investing in land in multiple states for the purposes of a larger investment portfolio.
When a company’s AZCC registration listed an Arizona address but other indicators suggested out-of-state ownership, reporters made a judgment call based on the strength of available evidence. We also considered whether a company’s operations were chiefly focused on providing local benefits—via locally sold produce, for example, or deep community roots—or designed to extract value elsewhere, such as through export-oriented production.
OpenCorporates was the first port of call in most instances, since it aggregates information from corporate filings in every state. Websites like OpenCorporates have archives of data on corporate registrations that help connect members and managers to other companies or addresses, as well as other details that can help identify shared ownership.
However, additional verification was often needed. Reporters cross-referenced open-source materials such as company websites, social media pages, court filings, public comments, farm subsidy records, licensing documents, news coverage, lectures and obituaries. Satellite imagery and mapping tools were also used to corroborate on-the-ground operations.
Because Limited Liability Corporations (LLCs) can legally own land and are designed to shield their owners, further backgrounding was critical in uncovering true ownership within the basin. The relative ease of creating LLCs also meant that a single property could be linked to multiple companies within the data, often organized under larger holding entities.
When summarizing the owners of the basin, we grouped together LLCs that we could reasonably prove are owned by the same parent company or individual, treating them as a single entity for the purposes of our analysis.
How can you show the movement of companies into the basin?
In Arizona’s rural areas, beyond the “reasonable use” standard for groundwater, the only requirement is filing an application with ADWR to drill a well. These applications contain details such as a well’s expected pumping volume and primary uses, forming the basis of the Well55 dataset. Using this data, we can see exactly when and roughly where companies showed interest in a region’s groundwater.
For each unique well application, there is a host of information available to the public, including the applicant’s name and state of residence. There are also three dates for each unique well ID: Application Date, Approval Date and Completion Date. We opted to base the bulk of our analysis on the Application Date field, as it marks the moment operators first sought to use water in the basin and is by far the most complete.
To get a clearer picture of successful well applications—and, therefore, the major groundwater users in the Hualapai basin—AZCIR looked at large-scale wells that listed irrigation as a possible use, excluding cancelled applications and abandoned or destroyed wells. We decided to include capped wells, as they could be uncapped and begin pumping again at any time. We then sorted results by the applicant’s state of residence.
From there, we cross-checked company names with names of parcel owners before doing additional backgrounding. Using the same techniques described earlier, reporters identified instances where multiple LLCs or individuals were linked to the same operation, even if the names did not directly match parcel owners. This process revealed that the number of key players in the basin was much smaller than the raw data initially suggested.
We also checked each well against registration codes to determine whether it was newly drilled or had been acquired from a previous owner.
Because the Well55 dataset is incomplete, AZCIR frequently compared findings against the original scanned records. Though forms must be submitted to ADWR as part of the well application and drilling processes, this does not always happen. Completion rates for certain fields, such as approval or installation dates, drop predictably and dramatically after the initial application stage, which shows a 99% completion rate. Missing forms often lack critical information like well depth and pump capacity, making it difficult to identify detailed statistical trends in areas with fewer wells.
