Subscribe to Upstream Tech’s California HydroForecast Outlook
Over the past century California constructed the nation’s largest state-built system of dams, aqueducts, and pumps to deliver water to its cities and farmers. While this engineered infrastructure has played a crucial role in supporting the country’s most populous state and most productive agricultural region (the Central Valley), climate change will cause extreme highs and lows in water flows beyond what California’s water infrastructure was built for. And, as the climate becomes more unpredictable, historical records of water flows will become less useful reference points for predicting water supply.
As a result, water managers are looking beyond the concrete to explore new tools to allocate water and manage shortages. Water markets are one promising mechanism. New sources of water and weather data, if properly applied, can further improve their function to help organizations manage their water risk.
At Upstream Tech, we’re motivated by our work to improve forecasting skill and apply it to the world’s pressing environmental issues. In collaboration with NGOs, state agencies, and water resource engineers, we’ve used our HydroForecast application to improve hydrologic understanding of the American West.
Drawing from this work, in this blog we explore how hydrologic information can better inform water market index values in California. As a part of this effort, we’re publishing a new newsletter that provides monthly updates on hydrologic conditions and forecasts in California’s most important reservoirs. Subscribe here.
Boosting Market Participation
In the last twenty years, water markets have emerged as a tool to help shift water use to where it is most valuable. A water market is like a stock market for water, where sellers lease or sell a legal right to use water to buyers that need it. For example, a farmer who fallows their fields during a drought might sell their water rights to another farmer who needs to irrigate their permanent crops.
In California, water right holders and traders purchase and sell water leases and permanent sales in an active market. In the last few years, two additions to this market sought to improve its transparency and to make participation smoother. In 2018, the Nasdaq Veles California Water Index (NQH2O) was created to track the spot rate price of water rights in California. In 2019, Nasdaq also partnered with Veles Water and WestWater Research to launch the Nasdaq Veles Water Index futures, which tracks the NQH2O index. Their futures contracts make it easier to participate in this highly complex market without incurring high transaction costs. While contracts are open for up to two years ahead, almost all transactions currently focus on the shortest futures contracts.
How Hydrologic Information Can Inform Index Price
Known for its big reservoir systems, much of California’s water storage is allocated through the Central Valley Project (CVP) and the State Water Project (SWP)’s 17 reservoirs. To date, market participants have typically relied on a patchwork of sources to guide their buying and selling of NQH2O futures. These include, but are not limited to, organizational or personal experience, conventions derived from historical trends, and news reports on water availability throughout the state.
Real-time hydrologic data can improve participant knowledge by providing standardized information that correlates with index price. By understanding how inflows into these reservoirs correlate with the NQH2O prices, we can identify the value of hydrologic information for market participants. Figure 1 provides a map of the basins that account for 80% of the inflows into the CVP and SWP: Shasta, Oroville, Folsom, New Melones, and Millerton.
Figure 2 illustrates the average annual observed inflow rate for each of these five reservoirs from 2000-2021. The patterns exhibit a signature spring snowmelt pulse, which occurs as the sun warms the mountain snowpack and generates flow into the reservoirs. This spring pulse is the most critical component in setting supply expectations for the year, as it dictates how much water is available for agriculture, environmental, municipal, and a range of other uses. Note how the spring pulse occurs earlier (March-May) among the big three reservoirs - Shasta, Oroville and Folsom - compared with New Melones and Millerton Lake, which receive inflow from further south in the Sierra Nevadas.
These inflow patterns align with changes (deltas) in the NQH2O index, as shown in Figure 3. This time series illustrates how much the index price changed on average in each month between 2013 and 2021. Most of the changes in the index (both up and down) occur from February to March and April to May with very little change in June, October, and the end of the year.
As Figures 2 and 3 illustrate, the timing of the largest increases and decreases in index price occur at the same time as the rise and fall of the spring snowmelt pulse. Investigation of observed inflow volumes during this spring pulse - from February to April - indicates a strong relationship to the next month’s index change. Further driving this point, a robust linear regression analysis resulted in an R2 of 0.35, with all coefficients showing statistical significance.
This suggests that we’re seeing a correlation between index price and the spring snowmelt pulse: hydrologic information consistently accounts for roughly a third of the future index price change during this period of the year. Again, the snowmelt season in California is the most important time of the year for determining water availability for human consumption, ecosystems, and agriculture.
Further, as streamflow patterns diverge from the historical average, the index value changes. Visualizing the anomaly, which is the difference between the observed inflow and the long-term average, demonstrates how the index value changes in unusually wet and dry periods. Figure 4 illustrates this relationship between the inflow anomaly and the NQH2O index from 2013-2021.
Generally, the index price goes higher during dry times and lower during wet times. This stands to reason: water is relatively more valuable when it is scarce. The index price shifts often trail peaks in inflow, and the largest price shifts typically occur after anomalies in flows. Each spike in price (2014, 2015, 2018, 2020, and 2021) follows an anomalously dry period, especially following lower than usual spring melts. In wet years (2017 and 2019), the index price dropped.
Next Steps: Using HydroForecast to Improve Market Efficiency
The strength of the relationship between index price and observed (current) inflows suggests that forecasts of the 1, 2, and 3-months ahead inflow forecasts would provide market participants with important foresight into the spring’s likelihood of being a ‘wetter’ or ‘drier’ year.
To put these insights into action, beginning at the start of 2022, Upstream Tech will distribute a monthly outlook report to support California water users. Subscribe Here.
The newsletter will include current hydrologic conditions and forecasts for the next month’s inflow, as well as temperature and precipitation forecasts, for the five largest reservoirs in the state projects.
In a future post, we’ll detail more thoroughly how HydroForecast’s forward looking predictions can support market participants in making more informed decisions. HydroForecast takes a unique, theory-guided approach to hydrological modeling using artificial intelligence, remote sensing, and a behemoth of a data processing pipeline. Behind the tech buzzwords is the most accurate and flexible operational hydrological forecasting system operationally available. In short: we are the best at predicting how much water will flow through any river over the next 10 days, 3 months, and beyond.
Interested in more detailed data or further-into-the-future forecasts? Connect with us at email@example.com.