GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Plus Icon Minus Icon Picture icon Hamburger Icon Close Icon Sorted
Guest Blogs

Undamming Rivers Could Make Room for PV

It’s time to remove hydroelectric dams that produce modest amounts of power yet do enormous damage to rivers and fish

Dams cause environmental damage to rivers and fish but many produce only modest amounts of power, according to the authors. If the dams were removed, there would be lots of room for photovoltaic (PV) arrays. This is the Oakdale Dam near Monticello, Indiana, completed in 1925.
Image Credit: Jim Hammer under license from Flickr.

Hydroelectric power is often touted as clean energy, but this claim is true only in the narrow sense of not causing air pollution. In many places, such as the U.S. East Coast, hydroelectric dams have damaged the ecological integrity of nearly every major river and have decimated runs of migratory fish.

This need not continue. Our rivers can be liberated from their concrete shackles, while also continuing to produce electricity at the site of former hydropower dams. How might that occur? A confluence of factors — the aging of many dams, the advent of industrial-scale alternative energy sources, and increasing recognition of the failure of traditional engineering approaches to sustain migratory fish populations — raises fresh possibilities for large rivers to continue to help provide power and simultaneously to have their biological legacies restored.

The answer may lie in “sharing” our dammed rivers, and the concept is straightforward. Remove aging hydroelectric dams, many of which produce relatively small amounts of electricity and are soon up for relicensing. When waters recede, rivers will occupy only part of the newly exposed reservoir bottoms. Let’s use these areas as a home for utility-scale photovoltaic (PV) and wind power installations, and let’s employ the existing power line infrastructure to the dams to connect the new solar and wind power facilities to the grid.

This vision both keeps the electricity flowing from these former hydropower sites, while helping to resurrect once-abundant fish runs, as has recently happened in Maine.

Fish access has a dismal record

More than a half-century of modern attempts to allow fish to traverse what often are sequences of dams that block access to their historical spawning reaches in eastern U.S. rivers presents a dismal record. Highly unnatural conveyances such as fish ladders are often only marginally helpful to fish on their upstream spawning runs, which is one reason why some migratory fish runs have fallen as much as five orders of magnitude.

Take Atlantic salmon, a revered game and food fish that once may have numbered a half million in U.S. rivers. In 2014, fewer than 400 attempted to reach their New England spawning grounds. Such relic populations are often protected from harvest, yet are still not meaningfully restored.

No other action can bring ecological integrity back to rivers as effectively as dam removal. Yet such efforts may come at the cost of a loss of hydropower. And so what many hoped would be a precedent-setting breaching of the Edwards Dam on Maine’s Kennebec River in 1999 — which had yielded only 3.5 megawatts of power — has not been followed by the dismantling of other, higher-wattage dams on the East Coast.

Yet the efficacy of dam removal to restore migratory fish was shown in the Kennebec after the Edwards Dam fell; for the first time in more than a century and a half, alewives, a species of herring, were able to access an upriver tributary, the Sebasticook. Within just a few years the Sebasticook’s run of alewives swelled from nonexistent to almost three million, supporting scores of bald eagles and an “alewife festival” that celebrates the Sebasticook’s extraordinary renewal.

Start by breaching aging dams

In “sharing” a river more equitably between energy production and its ecological imperatives, the critical step would be the breaching of existing dams. Though that may seem improvident — if not downright radical — it is important to remember that many of these concrete walls are middle-aged or older and will be reaching their life expectancies in the coming decades. Deteriorating dams are a serious public safety concern — one likely to increase as climate change generates more frequent and intense storms.

We believe the compelling ecological and impending structural reasons for dam removals should be considered in light of the rapidly evolving national energy landscape, and that together they signal exciting possibilities for a dramatically improved stewardship of major rivers. Fortunately, traditional hydropower facilities already offer the real estate that lies under reservoirs and existing electrical transmission lines that could be used by renewable energy sources.

In breaching a dam and draining a reservoir, substantial areas of land could become available for new uses. Take the Conowingo Dam in Maryland, for example. The Conowingo is the largest of four hydroelectric dams on the lower 55 miles of the Susquehanna River and sits only nine miles above the head of Chesapeake Bay. Its 572-megawatt capacity is fed by a 9,000-acre reservoir that also serves as an emergency water supply for Baltimore, and provides water for cooling intakes at the nearby Peach Bottom nuclear plant. The pool is also used by recreational boaters and fishers.

If the Conowingo Dam were removed, this would free up more than enough area to replace the lost hydroelectric generation with power from solar parks along the former reservoir bottom, and to allow for other land uses, such as creation of fringing wetlands and forests.

For comparative scale, California’s new 392-megawatt Ivanpah Solar Electric Generating System has three units occupying 3,500 acres. More sun shines on the Mojave than in the mid-Atlantic region, but according to the National Renewable Energy Laboratory calculator, acre for acre, the Conowingo region should support 76 percent of the power-generating capacity of the desert. Thus, about three-quarters of the river bottom would need to be in solar to match the output of Ivanpah.

One other issue facing the Conowingo Dam removal would be the sediments behind the dam that would need to be stabilized. The reservoir itself is close to capacity, and current plans are to dredge the pool, at an estimated cost of $48 million to $267 million annually. Those who are concerned for the ecological health of the Chesapeake Bay fear that if the dam is removed, millions of tons of sediment, enriched with nutrients and (potentially) toxic substances, could pour into the bay. But sediment stabilization is routinely done in dam removals and could be safely accomplished with careful design and engineering.

What about pushback from local residents?

Finally, what of the pushback by those who cherish the status quo? Few local residents were alive when the Conowingo Reservoir began filling in 1928, so the big pool is their cultural heritage. Surely any such drastic change would be hotly debated in many forums. But only a small number of houses exist on the 29 miles of shoreline that would be affected if the reservoir were removed.

The same issues were faced in the debates about removing mainstem dams in the Penobscot River in Maine, and eventually a consensus emerged there. Preservation of power generation (diverted to smaller tributaries) was important to closing the deal, and will likely be important in other cases. And although man-made reservoirs have their aficionados, rivers often have more of them — the scores who appreciate the fishing, paddling, and nature watching they provide. One study showed large economic benefits from the Edwards Dam removal.

And what about the nuclear plant and Baltimore’s emergency water supply? The Peach Bottom plant could install water-miserly, closed-cycle cooling towers, and Baltimore could still withdraw water from the Susquehanna in an emergency.

There are other potential tools available to help share rivers. Any remaining backwater ponds could be outfitted with floating solar panel arrays, as used successfully in Japan. Also, because reservoirs are nestled in valleys, in some instances the surrounding ridges might host wind turbines. Though combined alternative energy sources such as these might alone make up or exceed the original hydropower lost, “run of the river” hydropower — in which only a portion of the current is routed through turbines — could also contribute. But, critically, while generating some hydropower, the river’s mainstem would remain free-flowing, opening the way for resurgent fish migrations.

On the Penobscot River, the precedent of restoring a major river while maintaining equivalency of energy production was recently accomplished. This was done by increasing hydroelectric generation capacity on a set of tributaries while reopening the mainstem channel through dam removals and more effective fishways — thus returning nearly 1,000 miles of river habitat to eleven species of sea-run fish, including Atlantic salmon, sturgeon, and river herring.

Other once biologically productive New England rivers now clogged with multiple dams — such as the Kennebec, Merrimack, Connecticut, and Housatonic — could be prime candidates for some of these new ways of thinking about the future of rivers.

We need a dramatically different vision

Other innovative approaches could also be explored. The previously submerged but newly available riverfront property might be sold or transferred for conservation easements or for parks or even environmentally sensitive residential development. The revenue from these sales could be used for solar or wind projects in other promising but underutilized locations, such as landfills and urban brown fields.

A discussion of new strategies is timely because we are about to double-down on the flawed status quo. The Federal Energy Regulatory Commission will be evaluating many East Coast hydro dams for relicensing within the next few years — licensing that would lock in the failed fish passage paradigm for as much as an additional half-century.

As two conservation biologists who study rivers, we believe it’s time to explore a dramatically different vision. It may be that hydro companies should not continue to act as the gatekeepers for what could otherwise be healthy rivers brimming with life. Certainly, society requires electrical power, and rivers already are part of our grid. The way forward just may be to share a river more equitably between renewable energy production and its natural ecology.

Karin Limburg is a professor of environmental biology at SUNY’s College of Environmental Science and Forestry. John Waldman is a Queens College professor of biology. This post was originally published at Yale Environment 360.


  1. srenia | | #1

    There are gateway dams for
    There are gateway dams for fish and other less costly solutions than getting rid of the dams. It should be looked at when applicable. Also concrete dams get stronger over time as they continue curing minus any structural issues. Lower square footage housing, less restriction on diesel fuel vehicles in the US and other easier solutions come to mind before dams. Where I live a good portion of electric comes from burning trash, wind farms and a dam. The article does make a strong case for keeping dams.

  2. user-1135248 | | #2

    For dams with decent production and relatively shallow reservoirs,
    with satisfactory migration support [or no documented need], why
    not go double-duty? Drain long enough to build a simple system of
    pylons to mount solar arrays on above the water. Or even just float
    such a structure without having to drain anything, and effectively
    recover the land area for solar *without* nuking the entire dam and all
    the work that went into its infrastructure in the first place. Simple
    substitution seems like rather wasteful thinking, e.g. a big step
    sideways instead of actual progress.


  3. Expert Member
    Dana Dorsett | | #3

    Utility scale solar on flood planes?
    There's a (not so) great idea for you! Covering the watershed areas feeding a freed up river with mega-acres of PV isn't exactly a benign thing for the environment, even if it COULD be flood protected. It's also a lousy investment in PV.

    The right place to put PV is near where the power is being used, relieving stress on the system, not "somewhere out on the grid" where it utilizes an order of magnitude more grid resources to deliver the power to the loads. While it's true that the installed $/watt COST of utility scale PV is lower than rooftop PV, it's also WORTH quite a bit less, since it requires a lot more grid infrastructure to make it useful.

    Floating utility scale PV arrays are being experimented with in flat-land constrained Japan, but it's not clear that covering lakes with floating PV makes any more sense than covering watersheds with PV in the US. ( )

    Utility scale solar currently holds the lion's share of PV in the US, but there's a vast amount of unexploited commercial building rooftop that makes far more grid sense and environmental sense to put to use before further screwing with reclaimed riverbanks, flood-planes, or puddles. Policy support at the utility rate structuring level could light a fuse under the commercial scale PV market, whether that PV is on the ratepayers side of the meter or the utility's side.

  4. charlie_sullivan | | #4

    PV needs hydropower intact or expanded
    As we increase the amount of PV on the grid, and bring it up to and beyond the amount of hydropower on the grid, it will get increasingly difficult to deal with the variation is PV output over time--between day and night and between sunny and cloudy days. Grid operators deal with variability all the time (usually primarily variations in load), by dispatching different generation sources, and now, increasingly controlling loads as well in "demand response" schemes. As the amount of PV increases, the need for fast-response dispatchable generation to complement it increases. The generation sources that are best at responding fast are gas turbines and hydroelectric. If we want to wean ourselves from fossil fuels, we will need hydro power more than ever.

    Nuclear is sometimes suggested as a good complement to PV and wind, but it is best operated at steady power--more so even than coal. It can't do the job of responding fast to compensate for variations in PV output and loads. Storage could be a "silver bullet" but getting batteries up to the scale to do the job of major hydro plants is not realistic in the short term. The realistic option there would be pumped hydro, and indeed a good option may be to add that capability to existing dams. Demand response will also need to be improved and used much more, but there are limits to what can be done with that.

    We are facing an major challenge in decarbonizing our electric power system. In order to do that, we need to muster all the resources we can. Hydroelectric systems not only provide low-carbon electricity, but they also facilitate the use of more PV and wind--their value to a low-carbon grid goes way beyond the megawatt hours they directly produce. Taking them down at this time when we most need them would be a tragic mistake.

    There's no point in trying to help fish by removing dams if we proceed, through climate change, to dry up some rivers, and change the temperature and salinity of the others.

    Once we have a carbon free electric grid with yet-to-be-invented storage solutions working with demand response to make it all work without any need to dispatch generation from hydro plants, we can have the luxury of considering removing dams.

  5. cussnu2 | | #5

    So the enviro nuts are
    So the enviro nuts are willing to trade fish lives for bird lives lost in windmill collisions. Things that make you go hmmmm.

  6. cussnu2 | | #6

    BTW, Hydro can be turned off and on pretty much at will
    How do you turn on solar at night or on a cloudy day? How do you turn on wind when there is none?

  7. Expert Member
    Dana Dorsett | | #7

    The "need" for more high dispatchables to support PV is a myth.
    The aggregate output curve of distributed PV in the day-ahead market is more predictable than the variability in load. The predictablity of that variability is very different from some random variability, and has proven to be far easier to manage than grid operators were thinking a few years ago (even areas where PV is sourcing at 110% of the mid-day load on some feeders.)

    In Hawaii the utility initially suspected the high concentration of PV in some feeder lines on Oahu for some of their ongoing grid disturbances, but one of the inverter vendors Enphase maintains huge data sets on a sub-minute time frame, and turned the data over to a third party for analysis. It was determined that not only was the high density of PV not a proximate cause, but that the PV was helping stabilizing the grid against those disturbances, and that with a software update to allowed longer ride-through the local PV was capable of stabilizing it even further. This was covered in some detail by Greentechmedia (which is down today, as they rework their website), and elsewhere. Enphase is of course crowing about it:

    Demand response is faster, cheaper and more targetable at points of grid congestion than any gas peaker or hygro generation could be. If FERC order 745 gets ruled on favorably by the Supremes, overturning the D.C. District Court, demand response will be able to bid into capacity markets, growth in demand response will grow an order of magnitude faster than it already is. (I was expecting that ruling sometime this summer, but according to my watch it's autumn already.)

    At the financial learning curve of grid battery technology (multiple types, capacities and response characteristics), batteries will be taking a far larger role in a variety of ancillary grid services such as frequency control & peak power generation, and do it better and more cheaply than fossil fired or hydro peaking. It's more realistic (even in the short term) than it seemed even a couple of years ago, and will CERTAINLY be cost effective by the time distributed PV hits levels where it's necessary to compensate for it. Those levels are being approached in parts of Hawaii right now- stay tuned. California's PUC is attempting to be proactive about it, and has mandated grid storage (on both sides of the meter) in the CAISO region to be able to manage anticipated fast-ramping needs in the future, but they may have overstated the actual need, given just how cheap automated demand response has become. Distributed grid-smart hot water heater elements are capable of quite a bit of load smoothing (even frequency control), and could, where the regulatory environment allowed aggregation, cash compensation, and grid-operator control over those assets.

    Younicos just won a contract in the ERCOT (Texas) in the past week or so to co-locate a grid battery operation at an existing smaller utility scale PV array. If batteries + PV can beat gas-peakers on price & performance in low-priced gas-rich Texas, you can bet it won't be long before it can beat it in your neighborhood.

    Batteries won't replace a Grand Coulee Dam very quickly, but distributed batteries could easily replace most of the small and mid-scale hydro in the northeastern US such as those discussed in the blog article and provide even better & more reliable grid stabilization by virtue of the faster response and being distributed on all sides of the substations. But they don't have to provide very much at all in the way of absolute capacity.

    Hydro isn't nearly as greenhouse gas free as it might seem. The methane production of lakes and lower carbon sequestration than many other uses (including forest) of that acreage gives them a greenhouse gas footprint, if not explicitly a carbon footprint. The total generation capacity of all the rinky-dink low-head New England hydro systems is miniscule compared to even one small nuke (like Vermont Yankee, now off line), let alone the ISO-NE system capacity as a whole, and very easy to compensate for. It's not a luxury to let them go, even the mid-sized versions. That capacity won't be missed by the grid operators unless all of it was taken down in very rapid fashion all at once. As a fraction of the total all the hydro in New England and the imported hydro from Canada adds up to less than 10% of the peak capacity and average annual power source for the ISO-NE grid, and is comparable to the aggregate of other utility scale renewables in the region.

    Behind-the meter renewables and cogenerators don't show up on ISO-NE's accounting, since it isn't directly monitored by the grid operator. But it's growth rate in the region as a whole has beaten all predictions, and could grow even faster if given better regulatory & policy support.

    Retiring mid-scale hydro like Maryland's Conowingo / Susquehanna dams would take some planning, but that wouldn't be particularly difficult in a single decade time frame. Massachusetts has added PV with roughly 2x the capacity of the Conowingo in less than a decade, if you're counting PV on either side of the meter, half of that in just the past 2 years. (And that's without covering flood-planes behind breached dams with PV. :-) It's much more widely and better-distributed than that. ) Maryland has barely scratched the surface of what's possible with PV, with currently less than 20% of the installed capacity in MA. But exponential growth could allow them to surpass MA by 2020, given sufficient policy support, even though the installed capacity in MA by 2020 will be something like 4-6x the current level (unless state policy takes a U-turn.)

    Bottom line, we really don't need small scale hydro going forward, but the last thing you'd want to do is use the recovered real-estate for ever more centralized utility scale PV. There are far more useful ways of deploying the PV. The fact that PV is so scalable makes it silly to site it at a remote valley rather than right at the load. The rationale for siting the dams in those locations was topographical, but has no bearing on what makes sense for PV. Just because you COULD install PV there doesn't mean you should.

  8. charlie_sullivan | | #8

    Can do without hydro, but better to retire fossil plants
    My subject line, "PV needs hydro" was an overstatement, as Dana explains. There are options including batteries, demand response and gas turbines. Demand response is in many ways the best option. But right now we are using gas turbines to do the job. I will be ready to consider dismantling dams once we replace the gas turbines with the combination of PV, demand response and batteries. The methane emissions associated with hydro are a real concern, but that is associated with new dams; with old ones it is water over the dam, so to speak. The opportunity to regrow forests in valleys could be important, but I doubt that alone justifies continued reliance on gas turbines.

  9. Expert Member
    Dana Dorsett | | #9

    The methane release is ongoing
    What, the algae stops blooming on old dam sites? Methane from artificial lakes is not a function of the age of the lake.

    Gas peakers are probably going away before small and mid-sized hydro, no matter how aggressively you tear them down.

    Gas peakers, are currently doing most of the ancillary grid stabilization, but in a growing number of locations wind farms are (and soon, storage &/or fast-ramping non-critical loads.) The ramp rates of small changes in rotor pitch on wind farms is faster and more accurate than peakers, and far cheaper to maintain & operate than spinning reserves. Xcel Energy uses wind farms this way every day, limiting the spinning reserve requirement, and load-tracks slower swings with combined cycle gas (at 2x the efficiency of a fast-ramping peaker) or other large scale power plants in exactly this way.

    It takes a tiny fraction off the annual capacity factor of the wind farms to do this, but it's economic even at recent years' record low prices for natural gas. The more wind power there is on a regional grid, the easier it is to pull this off, but it doesn't take a huge amount.

    While it's technically possible to do some ancillary services and load tracking with small scale hydro, it's not clear that it is in fact happening anywhere. Can anybody point to an instance where small hydro is being paid for grid stabilization services or load tracking in the northeast? (Somewhere in VT, maybe? Or maybe not.) I suspect only in the heavy-hydro TVA or BPA regions would you find hydro resources being used for load balancing, and nearly nowhere for ancillary services such as voltage optimization & frequency control.

    Grid aware hot water heaters have been proven as fast ramping (to the millisecond) loads, and it's only a matter of time before they are deployed, with the ancillary services paid for, undercutting the cost of stabilizing the grid with peakers. It's primarily a regulatory problem, not a cost or technical problem. There are just SO many much cheaper ways to stabilize the grid, with methods that are more precise and easier to site in problem areas than fast-ramping peakers, it seems clear that they don't have much shelf-life left. Only in areas where those capital investments are still rate-based will those generators be economic to operate after Y2030. They're uneconomic NOW in many areas, even before regulatory changes open up those markets to more players.

    The imminent demise of fossil fired peakers is being covered one of this afternoon's GreenTechMedia blog bits:

  10. charlie_sullivan | | #10

    Need a plan to get off fossil fuels
    Dana, there's a lot you say that I agree with--like that integrating more PV is not a problem, and that demand response is a great way to stabilize the grid. But we are a long way from a renewable grid without using fossil fuel to ramp. I am thinking on a 10-30 year time scale and hoping to get rid of most of the fossil fuels, not wondering how we can get PV up to 10% in the next 5 years. We could do that without hydro, but we need a suite of tools for sure, and it will be much cheaper if we use all the best tools. Hydro is in fact one of the best.

    Wind and solar can both do regulation on any time scale by throwing away energy. Yes, the amount lost can be small, if the amount you want to accomplish is small. Wind can also do shorter time scale stuff using the inertia of the turbine. That's all great. But it's not quite as great as hydro, which can do it without throwing away any energy. With hydro there's no need to confine the contribution to small perturbations--it can participate in the heavy lifting. And the dams in New England are in fact used for that. (They are also, by the way, higher capacity than VT Yankee which didn't do any of the load following anyway--VT Yankee was 0.64 GW of the 31 GW on ISO-NE. That's 2%, clearly smaller than the hydro capacity located in New England (4%), not counting imports (~4% I think) or pumped storage (5%). Although I'm not sure why that comparison was relevant.)

    As quick evidence that they are used that way, here's a page complaining about the effects of the peaking operation of a ~40 MW dam on the Connecticut River:
    Not enough data to show that it's any faster than hourly, but it's at least participating in the hourly market.

    The GreenTech story about peakers being on the way out is not about using fewer simple gas turbines and more combined cycle. Big efficiency improvement but still fossil fueled. And the story about algae growing based on farm runoff is alarming, but may not be representative of the typical reservoir--and that problem may be proportional to fertilizer input not water area. The conventional wisdom (which I admit may not be the full story) is that the methane release is the submerged forest decaying, and that's a one-time event, though one that takes something on the order of a decade.

  11. Wiredforstereo | | #11

    A great lack of information
    Many commenters on this issue are woefully misinformed. Many comments in this and other forums essentially consist of "No, surely we can't get rid of hydro, has anybody thought of some other way to move the fish?"

    But it goes much deeper than that. I live near the Rogue River in Oregon which has undergone a series of high profile dam removals of late. The fact is, the vast majority of dams that people campaign to remove are old, no longer serving their original purpose, low producing, and/or blocking fish unnecessarily. One dam recently removed here has stood unused for anything for more than 40 years. Yet it blocks fish, and recreation.

    Many people also don't realize that dams not only block fish going upstream, but also block and kill fish going downstream. Oftentimes, young fish are shredded going through turbines or die of heat stroke in the warm pool created by the dam.

    Furthermore, dams block natural sediment flow, which does even more to mess the fish up because of the loss of sand banks and gravel bars in which to spawn.

    Dams are not being added in large part, and won't be in the future. Hydro power is simply too costly, culturally and environmentally to make an economical choice. We have already reached the point where more dams are coming down than going up. Hydro was good for the primitive economy of the last few hundred years, but it's time we grow out of it. And it's only the right thing to do to clean up our mess on the way out. Fish like salmon, trout, and others that breed inland are a huge destroyed and wasted resource that will return after we get out of the way.

Log in or create an account to post a comment.



Recent Questions and Replies

  • |
  • |
  • |
  • |