Arizona’s Next Water Crisis

Category: Best data-driven reporting (small and large newsrooms)

Country/area: United States

Organisation: The Arizona Republic

Organisation size: Big

Publication date: 12 May 2019

Credit: Rob O’Dell, Ian James, Mark Henle

Project description:

Rural Arizona is the Wild West for water. In most desert areas, water deep underground is the only source available. Yet there are no rules limiting how many wells can be drilled, or how much water can be pumped. 

Big farming companies owned by out-of-state investors and foreign agriculture giants have descended on rural Arizona to drill massive wells, often shipping their crops out of the country. 

An Arizona Republic investigation showed this free-for-all is draining away the water that homeowners also depend on, leaving some with dry wells and several rural areas of Arizona facing a crisis.


Impact reached:

The Arizona Republic’s series is the most comprehensive examination ever done of Arizona’s groundwater, which accounts for 40 percent of the state’s water use.

The investigation found the water levels in nearly one in four wells in Arizona’s groundwater monitoring program have dropped more than 100 feet since they were drilled, a loss that scientists and water experts say is likely irrecoverable.

On top of that, the number of newly drilled wells has been accelerating, and the wells are hitting water at deeper levels. The conditions are worst in farming areas where there are no limits on pumping. And the problems have worsened with the arrival of big corporate farms and investors. 

The series unmasked for the first time several of the corporate farms and investment funds that are likely the biggest water users in the state. The Republic profiled a Saudi farm using Arizona resources to ship hay back home, a 37,000-acre dairy that grows almost all its own food for its cattle while drilling wells a half-mile deep, and an investment fund that is renting out cropland even though there isn’t enough water for the local high school to water its football field. 

The series has had an immediate impact. Members of the Arizona Legislature have announced six bills to strengthen groundwater rules in unregulated areas. Legislators have credited the Republic’s series for exposing the crisis in rural areas. Momentum appears to be building for reforms that would be the first substantial changes to the state’s groundwater law in 40 years, as the Republican Speaker of the House said legislation is needed after years of blocking water reform bills.


Techniques/technologies used:

The investigation involved an unprecedented analysis of Arizona water-level data for more than 33,000 wells throughout Arizona, including some records going back more than 100 years, and nearly 250,000 well-drilling records.

We used a combination of MySql, Excel, Python to analyze the two datasets, which were very different in scope and detail. Much of the analysis was done in MySql. Python was used for some data work where MySql didn’t work as well, such as rank functions. Much of the MySql data work consisted of subqueries where we examined well data by groundwater basin, grouping the well data by together basin and across various time periods. 

ArcGIS was used to analyze property records, which the Republic obtained from half of Arizona’s counties. We were able to use the property records and well-drilling data — which often included LLCs — to find the owners of the massive corporate farms and determine how much farmland they owned and where it was located.


What was the hardest part of this project?

The main difficulty with the data was with the 33,000 well monitoring records because some of the dataset was limited in scope. There was ample data available for some wells, with records going back more than 100 years. But for other wells, the data was spotty and sometimes there were only a few readings over decades. Also, a limiting factor was that the program was completely voluntary, meaning there was no data for hundreds of thousands of wells. We got around this by aggregating the data by groundwater basin and using five and 10 year periods to compare the basins so there was enough relevant data.

The well drilling records were problematic for a different reason. The records were part of a large relational database and because the way the legacy Oracle database was set up, we had to do extensive cleaning and data work to create the correct results. There was a main table that had the latest regulatory action and a historical table that had all the regulatory history, but without the most recent action. We needed to combine the two. 

There was no obvious smoking gun in the well monitoring data. Because the well data was voluntary and sporadically measured, we had to look over long periods — in places over decades —  to show the effect that pumping has had on aquifers. Some of the analysis involved layering the data alongside examples of large farms we had discovered through our reporting in rural areas of Arizona. 

Through this unprecedented statewide analysis of groundwater data, we were able to show how unregulated pumping of groundwater is draining away the only water supply much of the state will ever have, as corporate megafarms continue to drain aquifers for quick profits.


What can others learn from this project?

This project began with a set of simple questions: How much have groundwater levels declined in Arizona in recent years? And have the water levels dropped more in unregulated parts of the state where there are no limits on pumping? 

Our reporting for this series demonstrated how data work can be the backbone of a big investigative project without the numbers themselves being front-and-center in the telling of the story. 

We were able to use the data as the main structure that proved the thesis of our series, but didn’t make the numbers the central focus of the story. Instead, we were able to tell the stories through the homeowners whose wells had gone dry, the farm owners who are concerned about their future and the residents who fear their towns may soon be turned into dust. 

The six stories in the series, together with the data visualizations, attempted to distill very complicated science and data into concepts that were easier for readers to comprehend and understand.


Project links: