BY ALEX REEP
California’s agricultural sector is reliant on the increasingly unpredictable mega-storms, called atmospheric rivers, that can dump as much water as twenty Mississippi rivers. Climate change has created unprecedented extremes: in some years there may only be one atmospheric river, while in others California can be inundated by twenty. As a result, farmers may face one year of extreme drought and be forced to fallow their land, and in others their fields may be unprepared for a deluge of water. In order for players in the nation’s largest agricultural sector to prepare for the increasingly unpredictable threat of atmospheric rivers, it is vital that the accuracy of climatic predictions improves through the effective use and stewardship of large data sets.
Atmospheric Rivers
On average, atmospheric rivers transport more than double the flow of the Amazon River, making them the largest freshwater “rivers” on Earth. They are long, narrow, horizontal concentrated streams of moist air that often generate heavy precipitation. Water vapor carried from the tropics to the West Coast in these atmospheric rivers has become a key feature of California’s hydrological cycle. Visits from a few large atmospheric rivers a year means the difference between a dry water year (drought) and a wet water year. In the span of only ten days per year, atmospheric rivers supply 30 to 50 percent of the state’s rain and snow. Warmer air holds more moisture, so atmospheric rivers are projected to increase in size – and risk – as the climate changes. Increasingly rapid processions of atmospheric rivers may make month-long state-wide superstorms the norm for California.
Thus, knowing how many atmospheric rivers will hit California – in addition to their variance and range – is essential to the state’s emergency preparedness and water management. Duration, intensity, and orientation at landfall significantly affects impact. Atmospheric rivers bring more than water – they can generate extreme surface winds, such as those that felled trees and power lines across the San Francisco Bay Area in 2021. It’s essential to understand how strong winds are, or how long they’ll last, given that some storms blow over quickly, while others linger and wreak havoc. Presently, atmospheric rivers bring snow to the high Sierras, but rising temperatures are shifting snow to rain, which can melt the snow pack and send even more water to lower elevation towns. They already cause the majority of economic losses associated with flooding in the western United States, and flood damages have been shown to increase exponentially along with atmospheric river intensity. In the face of these challenges, it’s becoming increasingly vital to improve our capacity to use weather data to modulate the seasonal and sub-seasonal predictability of landfalling atmospheric rivers for water resource management and flood control. The disaster response plans, infrastructure, and hazard maps of the state’s past may soon be inadequate to address the flood risks of the future.
California’s Agricultural Sector
Changing trends in atmospheric rivers will determine the viability of California’s agricultural sector, which is the largest in the United States. California’s agriculture generates $50 billion per year and employs over 420,000 people. Water is a primary concern. Drought is a recurring feature in the state’s climate, and climate change has further constrained water availability and increased crop water demands, leading to chronic over-pumping and damaged infrastructure. During an average rainy season, the state may experience five or six atmospheric rivers. Other years, however, may experience only one atmospheric river, while the next year could experience twenty. Approximately three atmospheric rivers annually are required for the state to reach its average yearly rainfall. The increasingly volatile variability between wet and dry years, or even wet and dry periods within a year, – deemed “precipitation whiplash” – is complicating farmers’ capacity to plan their growing season and safeguard their crops.
As Steve Johnson, a private meteorologist for farmers in California, reports to California Ag Today, “The year [2021] was difficult because some of these storms – in fact the big ones – even though they showed up in the 14, 16-day period, they didn’t look gigantic until about day seven or eight. Then they started gaining and gaining and gaining, and by day four or day five they looked monstrous. Well, that’s not very much time to prepare.” Poor water management infrastructure poses a major risk for farmers, as flooding disrupts the supply chain and wipes out investments in seeds, soil, and other agricultural inputs. Atmospheric rivers may provide the water required to grow more vegetation, but that plant matter could generate wildfires if followed by a drought year. Farmers face immense pressure to better understand atmospheric rivers as key components of the state’s variable hydroclimate.
Atmospheric Rivers and Climate Change
Improving our understanding and predictability of landfalling atmospheric rivers is critical for making decisions about the state’s water supply – particularly in the face of climate change. The increasingly dramatic transitions from drought to extreme wet conditions are being more accurately evaluated by modern scientific advancements in modeling, and atmospheric rivers are notoriously difficult to model. For example, even if two atmospheric rivers look identical on a satellite, the origin of the droplets makes a big difference in precipitation yield. The nuclei centers of rain droplets can form from either ocean salt or man-made particles, like burnt fossil fuels. An atmospheric river that has a man-made seed particle, however, will generate 40% less precipitation than its salty, more-natural counterpart. The size and frequency of atmospheric rivers depends on changes in the sea surface temperatures across the Pacific Ocean, which varies significantly based on if it’s an El Niño year or time affected by even less-understood climatic phenomena. Researchers recently coupled characteristics of atmospheric rivers at landfall with higher emissions scenarios to predict how flood damages in the western U.S. will vary due to climate change, finding that damages could increase from $1 billion in the historical period to $2.3 billion in the 2090s with an intermediate emission scenario.
Research Gaps on Atmospheric Rivers
Despite the significant research into understanding atmospheric rivers, there is still much we don’t understand – leaving California’s agricultural sector at risk of devastating losses. Supercomputers can process big data to help leverage existing knowledge and improve the forecasting capacity of Atmospheric River Detection Tools (ARDTs). That being said, there are many gaps in preexisting tools. At the moment, forecasters are able to identify incoming atmospheric rivers up to a week in advance, but they are seldom able to identify where they will hit or how intense they will be. Understanding how atmospheric rivers move over land – and how these patterns will evolve as the climate changes – will allow state officials to improve water resource management and flood protection for the benefit of agriculture, recreation, industry, municipal needs, and ecosystems.
Big Data Use for Sustainable Transitions
Agricultural Big Data (AgBD) is an essential component of refining long-range weather forecasts and predictive modeling for a changing climate. AgBD from observations of crop development and stress on multiple farms over large regions and time scales can be pooled and mined to assess relationships between site characteristics, weather, and crop performance in order to customize management conditions. Big data is a powerful tool for climatic projections and planning, but data alone is not a panacea for food systems problems. An important consideration is what data is being fed into these models, who has access to data and models, and which farms have the means to prepare and plan for climatic phenomena, like atmospheric rivers.
Although various types of AgBD have been made publicly available by providers, such as the U.S. Department of Agriculture, the U.S. Bureau of Labor Statistics, the National Oceanic and Atmospheric Administration (NOAA), and the National Aeronautics and Space Administration (NASA), not all valuable datasets are open to the public. Private AgBD includes privately held data collected by agricultural companies, financial institutions and individual farmers who keep data for internal use only due to privacy or business concerns, such as competition in commodity markets.
The Midwest Big Data Hub identified five primary limitations of current AgBD, including: (1) limited data storage and preservation, (2) data sharing barriers, (3) insufficient data documentation, (4) lack of connection between observation and theory, and (5) missing crucial data. Security concerns discourage private agricultural dataset sharing, inadequate documentation of metadata properties complicates its reuse, current empirical crop models cannot be extrapolated to new climate patterns, and water use data remains largely unavailable.
An additional shortcoming of AgBD is its capacity to reinforce the dominant agricultural regime. Artificial intelligence on commercial agriculture, for example, only includes data on a small selection of major agronomic crops, representing a design bias toward commodity crops and capital-intensive farms. As a result, AgBD often serves to entrench the market advantage of large agribusinesses, who collect and control the majority of data, which could cause the food system to become increasingly characterized by an industrial mode of agriculture.
Recommendations:
Comprehensive studies of various ARDTs have made it clear that no single tool can be recommended universally. The complexity of this evolving climatic phenomena necessitates a multimodal approach for water management, strategic agricultural planning, and disaster preparedness. Considerations of how agricultural big data can help or hinder transitions toward more sustainable food systems are required to prevent further entrenchment of the industrial regime.
1) Capturing Water from Atmospheric Rivers
Atmospheric rivers have great potential for replenishing aquifers and recharging groundwater stores. California must invest in infrastructure that directs heavy flows of water to underground aquifers, particularly in spaces vacant from mining operations. Improving groundwater bank storage will require upgraded storage and conveyance infrastructure, in addition to expedited permitting for recharge projects. Additionally, California’s urban and rural spaces require “nature-based” solutions to achieving water resiliency. In rural areas, farmers should consider how to improve field and forest health in order to improve soil permeability and yield for groundwater. In cities, planners should invest in green infrastructure that captures, cleans, and stores runoff water. On a governmental level, water users should be offered incentives to adopt water-saving technology, water entitlements must be reallocated based on environmental considerations, and drought plans should be drafted. Across the state, water usage data must be gathered to better understand who is using water, at what times, and in what quantities in order to improve demand management. Considerations of water circularity must be applied at every scale to preserve water resources for ecosystem health.
2) Using Data for Water Management
Big data must be applied responsibly to improve forecasting, water management planning, and seasonal forecasting. Improving data on watershed monitoring and modern weather and water forecasting will help water managers bolster their adoption of forecast-informed reservoir operations (FIRO) to selectively retain or release water from reservoirs according to current and forecasted conditions. Seasonal precipitation forecasts for California perform poorly, demonstrating the need to transition from the climate modes traditionally used in seasonal forecasts toward regional teleconnection patterns, which have proven to play a stronger role in precipitation variability. Extended range predictions can better condition forecasts by accounting for interactions between the climate state and regional climate modes, which plays a big role in determining how atmospheric rivers will make landfall. Finally, the future of planning for atmospheric rivers will rely on data from the past. Retrospective data spanning multiple decades can be used to test the accuracy of atmospheric river modeling tools in order to improve their sub-seasonal to seasonal predictability.
3) Planning for Climate Change
General Circulation Models (GCMs) are numerical models that represent physical processes in the atmosphere, ocean, cryosphere and land surface, and are the most advanced modern tools used to simulate the response of the global climate system to increasing greenhouse gas concentrations. Atmospheric river detection tools must be applied to GCMs to simulate how atmospheric rivers may change as the climate changes. Continued observational and modeling capability development, including those based on machine learning, is critical to refining ARDTs and identify which communities are most vulnerable to the climate-related intensification of atmospheric rivers. Furthermore, increased focus on county-level projections must be developed to identify highest-risk zones so that policymakers can target efforts to increase climate change resilience. GCM projections can be linked with historic records of flood damages to identify areas at future risk of flooding, which allows local and regional planners to target investments in climate adaptation – including managed retreat for individuals living in floodplains and support for farmers with at-risk fields.
4) Ensuring Agricultural Big Data Will Not Entrench the Industrial Regime
There is a need to expand curated cyber-infrastructure, along with an ongoing emphasis on digital democracy. Data centers with long-term storage will allow various agricultural actors to compile geographically-diverse datasets to support longitudinal studies on the impact of atmospheric rivers on food production. The expansion and development of software that promotes knowledge exchange between academia, government, industry, and NGOs, such as the National Science Foundation’s Big Data Regional Innovation Hubs (BD Hubs), will reduce barriers to data sharing, information access, and strategic application to allow farmers to prepare for phenomena, such as atmospheric rivers. Digital democracy calls for decision making power to be transferred from monopolistic power holders to a broader group of citizens who influence how data is collected, processed, and applied. Technological and data sovereignty thus implies agency to be fundamentally redistributed to affected stakeholders such that they are able to use AgBD to study and respond to the challenges they face.
Conclusion
Increasing anthropogenic effects on the global climatic system mean that lessons from past atmospheric rivers may no longer apply to those we should expect to experience in the future. Utilizing big data to study these storm systems helps us to understand the mechanics of atmospheric rivers in order to better forecast impacts, communicate warnings, and improve response. There is an urgent need to invest in climate projection technology and open data infrastructure in order to ensure that stakeholders at all levels of agricultural production and distribution can prepare for atmospheric rivers of increasingly variable size, duration, and impact. Regional and state level authorities should utilize projections from models to identify communities and farms at the highest-risk of flood damage or drought in order to invest in climate adaptation, including water storage infrastructure.
Comments