Open cloud platform delivers water consumption models

Water managers and farmers in the West will soon be able to more accurately track water consumption by crops and other vegetation through a web application that uses data from satellites and weather stations.

The Environmental Defense Fund (EDF), NASA, the Desert Research Institute and Google last month announced plans to build the platform -- called OpenET -- to help decision-makers better manage irrigation, develop more accurate budgets and support groundwater management.

The app gets its name by combining two elements: “Open” refers to the open data and collaborative development of the platform, while “ET” stands for evapotranspiration, the process by which water evaporates from the land and transpires from plants.

There are multiple ways to estimate ET, but it’s difficult and expensive to get accurate, timely satellite-based data on the amount of water used to grow food.  “Estimates from different approaches often differ and sometimes those differences can be large so the data is then fragmented,” said Robyn Grimm, senior manager for water information systems at EDF. “This can lead to confusion and a lack of trust in the data, and it keeps it out of the hands of many who could really stand to benefit from the information.”

OpenET is solving for that by bringing many of the leading scientists behind those methodologies together onto one team and performing an intercomparison of the data from those models to form a consensus view on what an ensemble estimate should look like, she said.

Here’s how it works: Algorithms take in publicly available datasets from satellites, weather stations and Google Earth, plus datasets from academic research teams, state and federal agencies and the 2008 USDA Common Land Unit database to bring together models for calculating how much water crops and vegetation consume. Because the process is data-intensive and expensive, EDF and its partners are working with Google Earth Engine, which provides a common cloud computing platform where all the models in OpenET can be run using consistent inputs, Grimm said.

By placing the individual algorithms and data for each model into Earth Engine, OpenET users can decide which model or models best address their water consumption information needs.

Another option is using OpenET’s ensemble model, which represents a consensus estimate for a specific location and time period.  For that approach, the OpenET team chose six of the most commonly used models to estimate ET and compared the models and estimates to other data, such as records on groundwater pumping and from 130 ground-based flux stations to get an accurate assessment across a variety of crop and land-cover types over time.

“In some cases, the ensemble value might be drawn from a single model. In others, it might be a weighted average of just a few models or of the full suite of models,” Grimm said. “Most importantly, it will represent a single value from the ensemble that is endorsed by many of the leading national experts in remote sensing of ET.”

HabitSeven, a web development firm, is building the interface for OpenET, which will be a responsive, web-based app with maps using vector and raster tiles through an open source geoserver to provide “a lightning fast, fully interactive experience,” she said. “OpenET pulls data from a database currently hosting over 20 billion data points and updates in real time. In addition to the web interface, a custom API is capable of providing data in milliseconds.”

OpenET has applications in several areas. At the field scale, it can help farmers and irrigators match irrigation to crop requirements and maximize “crop per drop” to improve conservation. For example, OpenET is working with a grower in Oregon’s Harney Basin to use ET data to reduce pumping by 15%, which saves up to 20% in energy costs, Oregon State Rep. Mark Owens said in a press statement. In Diamond Valley, Nev., farmers have experimented with new irrigation techniques and can use OpenET to determine if they were successful.

On a larger scale, the Rosedale-Rio Bravo Water Storage District in California is using OpenET data as a foundation of an accounting and trading platform. The general manager there likens it to giving water managers a view into their checking account to help them make decisions about managing their budgets and to connect buyers and sellers within the trading program, Grimm said.

“To comply with the new groundwater law in California, it’s imperative to have accurate, transparent water use data to serve to build a groundwater budget,” General Manager Eric Averett said in the release.

OpenET is expected to be available to the public in 2021. “Next year, when OpenET goes live, anyone will be able to go to the OpenET website and zoom in on any geographic region in 17 western states and get near real-time ET data for that location — down to the field scale,” Grimm said. OpenET plans to expand to the entire country and beyond, she added.

Before the platform launches, the team wants to finalize a custom reporting interface and application programming interface where users can privately define their geographic boundaries, time frames and other parameters for custom reports. A educational pages about water budgets, ET, consumptive water use and proper applications of ET data  are also in the works.