Merced RIver Trail; Photo by Steve Greenwood/BLM

WATER COMMISSION: Merced River Watershed Flood MAR Study

Kamyar Guivetchi, Manager of DWR’s Division of Planning has often referred to Flood Managed Aquifer Recharge (or Flood MAR) as a “moon shot” for recharging depleted groundwater basins, but just how much Flood MAR can contribute to groundwater recharge in a watershed is unknown.  However, the Department of Water Resources’ Integrated Watershed Management staff is underway with a pilot study to look at the potential for Flood MAR in the Merced River watershed. 

At the October meeting of the California Water Commission, Mr. Guivetchi and David Arrate, Senior Water Resources Engineer with the Department of Water Resources, gave a presentation on the study and shared some of the preliminary results.

Introduction to Flood MAR

Kamyar Guivetchi began by noting that as a Mediterranean climate, extreme events are a part of California water.  The state rarely experiences “average” water conditions and in some years, there are both floods and droughts.  Climate change is making those extremes more frequent and more consequential, so what is important to understand are that the consequences of climate change are accelerating and exacerbating these extreme events, and that means we have to transform the way we think about planning, managing, and operating our water management systems, he said.  And that means we need to go beyond good coordination and even collaboration to co-managing our water sectors at a watershed scale.

The slide shows a flower diagram with the petals of the flower being each of the major water sectors.  Historically and even today, the water sector intends to operate in the periphery of the petals where there’s little or no overlap between the water sectors, but if the state is to move toward the center of sustainability and resilience, this means the water sectors need to work much more closely together to get that overlap, he noted.

Integrated Watershed Management involves multi-sector collaboration for the purpose of bringing multiple disciplines together to plan multi-benefit projects and then knit together the various pots of money from these various water sectors to implement those projects.

Flood managed aquifer recharge, or Flood MAR,  in many ways, epitomizes integrated watershed management,” said Mr. Guivetchi.  “And in its simplest form, it’s about using higher flows, peak flows, for the purpose of recharging depleted aquifers, and doing it in many cases on agricultural lands, working landscapes, and natural managed lands.”

The state of California in a number of venues and initiatives has recommended Flood Managed Aquifer Recharge, including in the 2017 Central Valley Flood Protection Plan and the final Water Resilience Portfolio. 

Some of the features of Flood Managed Aquifer Recharge is that it is a voluntary, public-private multi-sector partnership that works with private landowners, public agencies, governments, and NGOs toward common purpose. 

Flood MAR is scalable; it’s been done on a small farm or area basis, but it can also be done at the GSA scale, the basin scale, or an entire watershed.  The larger the scale, the more benefits are accrued.  It’s multi-faceted; there’s not just one type of Flood MAR project.

Flood MAR is one of the untapped areas of California’s water portfolio,” said Mr. Guivetchi.  “Our watershed studies are looking at climate change vulnerability assessments and adaptation studies are taking a headwater to groundwater approach.  That could include using things like Forecast Informed Reservoir Operations to anticipate an oncoming atmospheric river, drawing down a reservoir in advance of that in order to catch more of that flood peak to reduce flood risk.  To do that, we’ll need new and expanded reservoir outlet works like they’ve done on the American River, and we’ll need more conveyance to be able to move that water from the reservoir into the working lands for Flood Managed Aquifer Recharge.”

“Part of this will involve identifying where the suitable recharge areas are, and then coming up with a compensation program and recharge credits to really incentivize the private participation in this program,” he continued.  “The better and more information we can get about suitable aquifers and recharge methods will help advance Flood MAR in each watershed.  Finally, we can do this in ways to provide restoration on ecosystems, both terrestrial and aquatic.”

In 2019, DWR partnered with Sustainable Conservation and other organizations to hold a Flood MAR forum where 200 participants over 2 days discussed the research and data development plan that was developed the prior year, and identified pilot projects and studies that can be done to fill the information gaps.  Participants also developed a number of recommendations that were organized under five goals or objectives. 

At the Department of Water Resources, they have focused Flood MAR activities in three areas:

  • Watershed studies such as the Merced and the Tuolumne rivers.
  • Pilot projects to fill information gaps identified in the R&D plan. The Department has identified a number of projects that can help demonstrate the efficacy or the effectiveness of Flood MAR and fill some of those data gaps. 
  • Developing guidance such as a white paper, a research & data development plan, and technical memos that describe watershed studies and future pilot projects.

Purpose of the study

Next, David Arrate, Senior Water Resources Engineer with the Department of Water Resources then gave a detailed presentation on the Merced River watershed Flood MAR reconnaissance study, which the Department is working on in collaboration with the Merced Irrigation District and with the help of consultants.

The purpose and goals of the study:

  • Proof of concept study to investigate Flood MAR concepts at the watershed scale. Flood MAR has been done for a while but typically at a small scale, so they are studying what kind of Flood MAR benefits there could be if it is scaled up to an entire watershed with the major reservoir, a major river, and a large conveyance network.
  • Integrated watershed modeling that integrates surface water, groundwater, flood, water supply, and all the different aspects of water resources management within a watershed through a total watershed model.
  • Assess multi-benefits, economics, and climate change.
  • Provide a template for future studies and projects. The Department is documenting everything and will be developing a series of technical memorandums on model integration, the lessons learned, and the study results.

Climate change

The study is looking at climate change using a decision-scaling methodology.  Typically most studies will consider only a handful of future scenarios, but with the decision scaling methodology, they can look at a range of potential future scenarios that cover what the Global Circulation Models (GCMs) show.  This study is considering 30 climate scenarios with temperature increase from 0 – 4 degrees and the range of mean annual precipitation from -20 to +30%. 

The graph on the slide shows the matrix of the different temperature and precipitation that will be modeled.  The modeling includes 100 years of hydrology from 1900 to 1999 that is simulated and based on historical gauge data.  From this, a system response curve can be generated, which is a gradient curve that shows what happens to the system under all the different climate change scenarios.

For example, we look at average annual inflow into Lake McClure, the reservoir on the Merced River,” said Mr. Arrate.  “The historic current condition is about 1 MAF average annual, and how does that change with all these potential climate scenarios.  This is pretty straightforward; you increase precipitation – more volume, decrease precipitation – less volume.  The increase in temperature also has some effect; you get a little less volume as you increase the temperature, so we can generate these types of graphics for all kinds of metrics – flood peaks, flood damages, supply, as part of the process we’re going through.”

Baseline and Flood MAR implementation levels

There are four steps to the study:

Baseline: The first step is to conduct a baseline climate vulnerability analysis of the watershed for baseline conditions, looking at how a changing climate will affect floods, surface water, reservoir operations, and groundwater.

Level 1: The first level of implementation is to consider the existing infrastructure and existing operations of the system, and how Flood MAR could work within existing conditions to skim water off of the system and move it through the conveyance network to do Flood MAR recharge.

Level 2: The second level of implementation considers reservoir reoperation with the existing infrastructure, such as Forecast Informed Reservoir Operations or an extended recharge pool during flood operations where water can be moved out of the reservoir and into the ground to create benefits that way.

Level 3:  The third level of implementation would consider new or expanded infrastructure, such as addressing bottlenecks within the conveyance system.  For example, some of the check structures and some of the lateral canals have pinch points where they are not allowing enough water to get to places that can take it, so those might be able to be expanded or even the whole conveyance system to get more water through there.

Mr. Arrate noted that this gets to the scalability of Flood MAR.  “Flood MAR is not just a one single project idea that you go out and implement,” he said.  “You can really scale it as you go along.  You can start doing level 1 things early on while you’re planning out what your reservoir operations might be, and then infrastructure always takes a long time to plan and construct, so those could be more long-term.  For example, Merced Irrigation District could implement all three levels over time, especially with climate change happening, they could gear level 2 and level 3 type things for more future climate scenarios.

Preliminary Results: Level One

Mr. Arrate then turned to the preliminary results of the study, beginning with an example from the climate vulnerability analysis of a peak flow on the Merced River of the 1956 flood event.  The green line shows the baseline condition in 1956 where the peak flow reached about 6000 cfs, which is about the channel capacity for the Merced River.  The yellow shows the same event but with +1 degree Celsius of climate change and a 10% increase in precipitation; peak flows are now more than double the channel capacity of 15,000 cfs and the length of the event is extended by about 3 weeks.  The red line shows that with +3 degrees Celsius and 10% precipitation, the peak event is now over 40,000 cfs.

So you really see how climate change has a large effect on this system,” he said.  “With just a little bit of warming and a little bit of precipitation, we see these much larger peaks that we’ve never really seen on this system before.”

The Department has been working with the Level 1 scenarios so far, and within the level, there are three scenarios:  initial, intermediate, and robust. 

The initial scenario is based on the State Water Board’s streamlined permitting guidelines, so the triggers are the daily 90th percent flow, the diversion amount is about 20% of total flow, and the time frame from December to March.  For the recharge, this scenario focuses on using the conveyance network which is comprised of unlined canals that seep at a fairly good rate, so they basically turn the conveyance network into a recharge basin.

The intermediate scenario expands on the initial scenario by loosening up some of the triggers and adding another month to the time frame, as well as adding on-farm recharge to the recharge through the canal system.

The robust scenario is the maximum book end within level 1 where they considered anything above minimum stream flow requirements with a small buffer above that.  While it might not be a realistic scenario, it’s a way to give a bookend on level 1, he said.

Mr. Arrate pointed out that Delta conditions are one of the triggers for Flood MAR operations; Flood Mar is only triggered when the Delta is in excess conditions.  He also noted that Level 1 analysis is a passive operation; they are fitting Flood MAR around existing operations and existing infrastructure to the maximum extent to assess what the benefits would be.  The level 2 and 3 analysis will consider reservoir reoperations and infrastructure changes and at that point, they can try to target certain benefits or other things.

Level 1 intermediate scenario: Water available for recharge

He presented a slide showing the baseline 100 years of hydrology without climate change events.  The bar graph shows how much water the water available for recharge (or WAFR) is based on the triggers.  Over 100 years, the model found 4.6 MAF available for recharge or an average of 46,500 per year, although it is highly variable – some years have over 250,000 acre-feet of water available for recharge; other years are down to just a few hundred or a few thousand. 

They next ran that through the models to figure out how much of that can be recharged.  The canals can recharge about 2 MAF.  Recharge basins recharge a very small amount, about 1% of that.  On-farm using agricultural fields uses 2.2 MAF. 

In the end, we do have about 374,000 acre-feet of unused water available for recharge (WAFR) in this scenario, most likely due to constriction points in the conveyance system or maybe the timing isn’t working with the ag lands, but for the most part, we end up recharging 92% of the water that was available for recharge.  Total recharge was a little over 4.2 MAF, average annual 43,000 acre-feet per year.”

Level 1 intermediate scenario: Groundwater recharge

What does that mean for the groundwater system?  Mr. Arrate presented a chart showing the baseline water budget and 100 years of inflows and outflows into the groundwater system.  The average annual baseline overdraft is about 49,000 acre-feet per year as determined through the simulations.

The orange bars show the additional flows from Flood MAR which is about 43,000 acre-feet of recharge per year on average, but overdraft has decreased by only about 17,000 acre-feet per year.  The remaining 26,000 acre-feet goes back into the streams, the Merced or San Joaquin River, and the neighboring subbasins.

The aquifers are all connected underneath, so even though we’re looking at our watershed, the neighboring watersheds’ aquifers connect to ours so some of that water is going over to other nearby subbasins,” said Mr. Arrate.  “In the end, Flood MAR is having an effect on the groundwater storage system within the Merced subbasin here as we’ve lessened the overdraft.”

The Integrated Models

Mr. Arrate then gave an overview of the models and how they all interconnect.  The study area is the Merced River which is outlined in orange and includes the purple shaded portion.

They go beyond their study area and look at the neighboring subbasins because they are all interconnected, he said.  “What happens in the subbasins outside our study area affects our study area and what we do within our study area with Flood MAR affects basins outside of it, so we go out and beyond for the groundwater.”

The model also considers Delta conditions as Flood MAR is triggered when the Delta is in excess, so there is modeling to see what the conditions in the Delta are.  The upper watershed feeds information into the Delta flows and other outer areas, and it’s also where the climate change vulnerability aspect such as changes in temperature and precipitation are factored in.

The flows in the upper watershed flow into Lake McClure, and all the water operations are incorporated into the model.  There are a number of creeks within our study area, so those were modeling along with any reservoirs on those creeks.  All the surface water interacts with the groundwater modeling and the information is passed back and forth between those models so anything that happens in the surface water system gets accounted for in the groundwater system, and things from groundwater modeling such as agricultural demands go back to the surface water models so those are integrated and iterative in their looping.

To look at Flood MAR, they used the Groundwater Recharge Assessment Tool developed by Sustainable Conservation and Earth Genome to take the water available for recharge and figure out how much can be actually recharged. 

Mr. Arrate said that there’s a lot of information that goes into this model.  First, the model looks at the district service area, the conveyance network, where all the water can be delivered within that area.  It uses a suitability index for the soils for recharge to target for Flood MAR with green being excellent, yellow is good, and orange is moderately good.  The model considers depth to groundwater, the depth to the Corcoran Clay, and the subsurface permeability as well.

The model considers land use; certain types of crops do well with Flood MAR so those crops are also programmed into the model.  Each crop reacts in a different way; they can be saturated for a certain amount of time, they have to dry out for a certain amount of time, and each crop has different cultural practices, so all of that is factored into the model.

Eventually what we get is this spatial list of potential recharge where can we actually do the recharge with the water is available,” said Mr. Arrate.  “So we get the water that’s available, we run it through this model, it actually does the recharge, and that information then goes into our groundwater model which we can see through the groundwater budget results what happens in the groundwater system.”

The last part of the model is flood modeling on the Merced River and some of the creeks around the city of Merced.  The model has structure data and other data pulled from various sources so for the climate vulnerability analysis, they can consider how future climate scenarios potentially have higher flood damages and then how can Flood MAR maybe mitigate some of those damages.


They are now working on the metrics, trying to figure out what metrics to use to quantify the climate vulnerability analysis and the Flood MAR benefits, such as flood risk, water supply both for surface water and groundwater, ecosystem management, vulnerability to climate change, and the effectiveness of Flood MAR strategies.

They have been working on building summary dashboards.  Some examples are shown below. 

As they work through the baseline and the level scenarios, they will be able to fill in with the absolute values of the scenarios and their relative change to the baseline.

The slide shows the results for climate change.  For flood risk, the peak flows are shown for baseline, 2040, and 2070 (#1).  Probabilities can be assigned to those climate scenarios, so they take the whole matrix and bring it down to single values, so for the 100 year peak flow, it goes from 6000 cfs to 12,000 to 19,000 for those two future projections.

The runoff from the upper watershed doesn’t change as much but the timing of the runoff is shifting (#2).  Notably, November 1st to March 31st shows an increase in runoff and April 1st to October 30th shows less runoff.

With warmer temperatures, agricultural demands increase so that is factored in (#3).  Surface water deliveries are factored in (#4) as are groundwater demands (#5).  So with the impact of surface water deliveries on storage and having less runoff later in the year, there is less carryover storage heading into the end of the year (#6).

Groundwater storage is also getting depleted faster as agricultural demands are going up, the reservoir doesn’t have quite as much storage so now the demands on the groundwater basin are going up.  The overdraft in the baseline condition was about 49,000 acre-feet per year; but at 2040, it is 80,000 acre-feet, and in 2070, over 100,000 acre-feet per year.

Mr. Arrate said the baseline analysis is underway, and they are still working on levels 2 and 3.  They hope to put some out of our tech memos towards the end of this year or early next year on the climate vulnerability analysis and on the level 1 analysis.

Public Comment

During the public comment period, Justin Fredrickson, Environmental Policy Analyst at the California Farm Bureau, pointed out that the study does not consider SGMA nor the unimpaired flows that were adopted by the State Water Board that have yet to be implemented.

When you look at the results of the 100 year modeling, they are exciting and it’s amazing technical work.  There’s a certain amount of water that erases in theory most of the overdraft potentially if everything hits just right, but once you add in those other elephants in the room, I’m afraid its illusory, it just goes away,” he said.  “The other is the Flood Board and DWR’s Flood SAFE department and I don’t know if the Army Corps comes in at all, but the flood implications of the three-fold increase are also quite dire.  This is where I get a little bit frustrated with arguments that there’s nothing we can do to increase water supply or adapt in anyway, so no surface water storage, new or modified, no conveyance, no new recharge, no underground storage, and no preparing for the future, that’s a policy choice.  If that’s our default, then that choice is we’re sending our agriculture elsewhere.  And I think that’s troubling and something that the Commission and the state of California need to grapple with.”

Print Friendly, PDF & Email