subscribe: Posts | Comments

Scientists differ on potential for river to save coast


Seven decades of subsidence, sediment starvation and channel dredging

Figure 1. Seventeen decades of subsidence, sediment starvation and channel dredging

Editor’s note: For nearly twenty years I’ve repeated a stock glib answer to the question of what can realistically be done to save and sustain much of Louisiana’s threatened coast. That answer has been that we “simply” need to reconnect the river to the delta, a concept that has become almost a truism, based on faith that “turning the river loose” would allow Mother Nature to reverse years of mismanagement. I no longer repeat that mantra, at least not without considerable hedging.

A fundamental premise of the ecosystem restoration component of Louisiana’s coastal protection and restoration program is that releasing most of the flow of the lower Mississippi River upstream from head-of-passes could offset projected landloss in the Pontchartrain and Barataria basins throughout the twenty-first century. Landloss in the Terrebonne basin would require more input from the Atchafalaya River.

What other solutions are out there? To be perfectly frank, there is no plan B. In my opinion, pumping dredged sediments through pipelines will be an important restoration tool in the near term (at least in critical local areas) but the Louisiana coast and other deltas around the world cannot be sustained on technology and brute force that depends on fossil fuel.

2090 projection by Blum and Roberts

Figure 2. 2090 projection by Blum and Roberts

Inundation of the emergent landscape can be envisioned as the relentless progressive expansion of  the tidal prism, the volume of water that regularly moves into and out of the coast, driven by astronomical and meteorological tides. The tidal prism grows larger each year by a volume precisely equal to the volume of mineral and organic sediment lost as the landscape sinks below sea level (Figure 1).

Neutralizing this process to achieve no net land loss would obviously require adding new mineral sediments and/or organic matter from net plant production at the same rate that relative sea level rise occurs and the volume of the tidal prism expands. This is not a trivial challenge.

In June 2009 LaCoastPost reported on a new study in Nature Geoscience on the capability of the Mississippi river to sustain its delta. The research was carried out by coastal geologists and Mississippi delta experts Mike Blum and Harry Roberts. Their results carry extremely sobering implications for Louisiana’s coastal protection and restoration program.

Blum and Roberts created a mass balance sediment budget for the deltaic plain. Their goal was to estimate the volume of sediment that would be required each year to offset estimated projected land loss, as extrapolated from historical data and different projections of accelerated sea level rise. Figure 2 illustrates their projection of the coast by the year 2090, under a no action scenario. The following is a quote from their study summary:

Sustaining existing delta surface area would require 18–24 billion tons of sediment, which is significantly more than can be drawn from the Mississippi River in its current state.* We conclude that significant drowning is inevitable, even if sediment loads are restored, because sea level is now rising at least three times faster than during delta-plain construction.

The annual input* of 18-24 million billion tons of sediment estimated by Blum and Roberts is 3.6 to 4.8 3,600-4,800 Superdome equivalents (SDEs). Dams on the Missouri River now trap half the sediment load formerly carried by the river. Thus these experts concluded that the triple whammy of delta subsidence, increasing sea level rise and a drastic reduction in riverborne sediments make the “salvation” of south Louisiana problematic at best and dubious at worst.

Figure 3. 2110 projection by Mohrig and Kim

Figure 3. 2110 projection by Mohrig and Kim

In striking contrast, on October 20 Eos (the journal of the American Geophysical Union) published the results of a separate study on a closely related topic. This research was also carried out by two highly credible geologists, David Mohrig and Wonsuck Kim, at the University of Texas (UT)-Austin. The latter effort was sponsored and summarized by the National Science Foundation (NSF) and also described by Science News.

Mohrig and Kim used a contrasting methodology and reached a far different and considerably more sanguine conclusion than did Blum and Roberts. Rather than constructing a mass balance budget for the entire delta, they used a numerical model to simulate twin east and west bank river diversions in Plaquemines Parish in the vicinity of the proposed Myrtle Grove diversion project, 93 mi south of New Orleans (Figure 3).

Their simulation model was calibrated using historical data on emergent land accretion of the Wax Lake Delta in Atchafalaya Bay, which formed as a result of the construction of the Wax Lake Outlet, a navigation channel dredged in 1942 off of the Atchafalaya River (Figure 4).

WaxLake subdelta lobe

Figure 4. WaxLake subdelta lobe

These scientists used their model to simulate the accretion of emergent landscape in the form of subdelta lobes into Barataria Bay on the west and Breton Sound (California Bay) on the east (Figure 3). The model was run under different estimates of diversion volumes (up to 45% of total river flow), suspended mineral sediment loads in the river and up to 4 mm/year sea level rise.  They concluded that this volume of flow could offset up to 45% of the projected land loss through the year 2110.

I’m curious and intrigued that the publication of these two important back-to-back studies on the crucial question re saving the coast, and studies with such contrasting results, has not generated serious public discussion. Both projections were run under similar conservative premises re rates of subsidence, sea level rise and river sediments.

So which study is more realistic?

Len Bahr (

*While writing this piece I flashed on the fact that, according to this reference, 159 million tons of suspended sediment (32 SDEs) exit the river mouth and fall off the Continental Shelf into deep water each year. This volume of mineral sediments would seem more than adequate to overcome the sediment deficit described by Blum and Roberts. Am I missing something here?

*As shown by the struck through words above I made a serious error in misquoting the sediment deficit reported by Blum and Roberts. As pointed out by David Muth, I used the figure 18-24 million tons per year, while the quote uses the figure 18-24 billion tons (no time period expressed).

Be Sociable, Share!
  1. I like the way you conduct your posts. Have a nice Thursday!

  2. I would move to Hawaii then travel the globe.

  3. Not meaning to be too critical, but in my opinion a LOT of similar events likely goes on in today’s fast paced computerized ‘scientific’ world.

    Someone grabs a number from somewhere, starts working with it, perhaps publishes their results, so someone else grabs these new results, uses it, etc…….

    A long chain of cards all perhaps based on incorrect initial bases/assumptions to begin with……

  4. Great interest in and confusion caused by this post has resulted in the administrative decision to post a followup, which is now under preparation. Stay tuned.

  5. I generally work with Pounds per Cubic Foot for specific weights.

    I converted 1.5 Billion Tons per Cubic Kilometer (per Mike Blum) to Pounds per Cubic Foot; and get about 85 Pounds per Cubic Foot.

    My experiences with fine materials such as silts and clays; which I assume you all are dealing with in the delta, is that their specific weights are a bit lower than this and are also quite variable.

    Perhaps on the order of 30-60 Pounds per Cubic Foot.

    For conversion, I used: 2000 Pounds per Ton; 0.6214 miles per kilometer; and 5280 feet per mile.

    I’ve also totally worked in fresh water; so ocean work might be different given the salinity of the water and the finer particle sizes.

  6. Mike Blum makes another IMPORTANT comment above; regarding SCALE. One group was looking at a SMALL area; and the other group at the DELTA scale. Its not easy to extrapolate/interpolate between the two.

    And I believe the same can be said for TIME scales.

    Sediment transport is an interesting beast; especially with the fine grain sizes you all are dealing with. There are so many variables that operate on a daily basis (tides, inflowing discharges, winds, etc) that in my mind its a real crap shoot to deal with given the chaotic nature of the water’s hydraulics interplaying with the fine grained sediments; and trying to extrapolate this to annual or decadal scales.

  7. Interesting. Is it Millions???? Is it Billions???? Is it??????

    Len, I hope you are starting to understand my previous comment about all the WRONG answers taken out to six decimal places…… AND why I question about anything anymore…..

    Hopefully the truth will emerge here and the innocent will be protected….. kinda like Dragnet…..

    • HH, David and Mike-
      I edited the text to correct my mistake of three orders of magnitude that David noticed. I also added the following footnote:
      “As shown by the struck through words above I made a serious error in misquoting the sediment deficit reported by Blum and Roberts. As pointed out by David Muth, I used the figure 18-24 million tons per year, while the quote uses the figure 18-24 billion tons (which, as Mike Blum has now pointed out, was intended to span roughly a century).”

      This problem with large numbers exemplifies why we need a visual crutch to understand the magnitude of millions and billions of ton of sediment – hence the recommendation of Superdome Equivalents or SDEs.

  8. David Muth says:


    Well one thing that the number from Blum and Roberts that you quote above is off and I did not notice the discrepancy. Their calculation is 18-24 Billion tons not Million Tons, which is a whole ‘nother matter. In fact, it means even a best case scenario– 100% capture, 20% organic inflation leaves us well short of the goal, assuming the “megatons” measure used by Anonymous as the figure from Kim et al is the same as Millions of Tons? Now I’m really confused and apolgize for sewing more confusion.

    I second Len’s request and would love to hear from the authors about what it all means. These are vitally important calculations. We need to know what we are up against and what we can realistically achieve.

    David Muth

  9. Several wide ranging values have been given here for ‘annual suspended sediment loads,’ which are generally presented in TONS per year.

    What values are being used, or are generally accepted, to convert the annual inflowing load in TONS into an annual sediment VOLUME; once the suspended sediment settles out???

    David Muth’s comment about things not being done at a STEADY RATE is right on!!! I think this is a part of the picture that is being totally missed and not properly accounted for.

    We definitely want our Kate and Edith too. On the one hand we want sediment to fall out everywhere it historically has; but on the other hand we don’t want water there also so we can habitate the land where Ole Man River used to call home occasionally…..

    • You bring up important points…….

      Suspended loads are measured daily, or averaged to give daily values. Blum and Roberts used values for the USGS Tarbert Landing site on the Mississippi, which is the first gauge below the Old River diversion, and in theory represents all the sediment available for diversions on the lower Mississippi farther downstream. There is some evidence that the numbers are 10% lower below Baton Rouge, which would mean storage in that reach. They also used the Simmesport data for the Atchafalaya. The digital period of record is 1975-present.

      The actual numbers are published, but the other point you raise… do you convert mass to volume…….that is also detailed in our paper, but it breaks down simply as follows:

      - sediment has a density of 2.7 grams or so / cubic centimeter
      - deposited sediment has a porosity of 40-50%, so density of a sediment plus pore mixture is about 1.5 g/cc
      - this translates to 1.5 BT per cubic kilometer

      Although everyone who works on these things would appreciate the need to get rid of “steady rates”, me included, it would not matter in this sort of modeling. It would be critical at small spatial scales, and over shorter time scales.

      With that said, as an example, sediment loads vary a lot, by 30% or more on an annual basis, so the only way to predict long term is to take the mean and some measure of the error. Same with all of the other variables. The primary mechanism for highly punctuated unsteadiness would be large storms, which will in almost all cases make things worse. So, not including them is a “best-case” scenario.

  10. All:

    I’m still missing something. If it takes “only” 18-24 MT/yr, and if in the worst case scenario the river has 105 MT/yr to give, even at 45% that’s 47.25MY/yr available, which gives us twice as much as the required 18-24 MT/yr. Right? Do Blum and Roberts use a much higher figure for RSL rise? Obviously the relationship between RSL rise and input is critical.

    It should also be obvious that we can’t settle for only 45% but must capture as close to 100% as possible (by decoupling the navigation system from the active distributaries); that we have to get the 65-70 MT/yr from the Atchafalaya working over a broader area (i.e. Terrebonne Parish); and that we need to build diversions far enough up the main stem to avoid the worst subsidence areas (i.e. above Port Sulphur). I’d also think it might be time to calculate the possible input to the system of annual and storm redistribution from an abandoned and retreating Bird’s Foot.

    I just heard at a conference that the real possibility exists that sea level won’t creep up at a steady and gradual pace of milimeters per year, but that it will periodically jump up at a pace closer to tens of centimeters a year as sections of continental ice give way. Punctuated equilibrium not steady rise. This is based upon studies of stranded sub-sea beaches on the Gulf continental shelf.

    David Muth

    • David-
      i’m also perplexed by what seems to be more sediment than necessary to keep the delta emergent if 100% of the flow is used. I encourage a member of each team, e.g., David Mohrig and Mike Blum, to provide a brief summary of how the two papers that seem to differ are in fact consistent.

      • I will take a shot. Sorry, but the specifics will make this sound like a long and windy road.

        Perhaps the first thing to do is clarify the numbers and units of measure.

        1. Blum and Roberts did not say 18-24 MT/yr, as in 18-24 million tons per year, but rather a total of 18-24 BT, as in 18-24 billion tons between now and 2100.

        2. The 18-24 BT total by year 2100 equates to 220-280 million tons per year, assuming diversions would start as early as 2020.

        3. If modern sediment load values of 135 MT/yr are taken then the total available comes to ~12 billion tons by the year 2100. If you add the Atchafalaya to that then you can get to about ~16.4 BT total. Each assumes diversions would be in place by 2020. If you use the most recent sediment load data, after the flood of 1993, they are less optimistic, and the Kim et al team uses smaller numbers of 118 million tons per year for the lower Mississippi, and 195 total for the Mississippi and Atchafalaya.

        And now, on the Kim et al. model from EOS…………

        The next issue is how much sediment will be diverted from the river to build new land. The number “45%” keeps appearing. That number is how much of the total load of the lower Mississippi the Kim et al. team simulated would be transferred through the diversions in their main model. So, 45% of the total means ~53 million tons available, because they use a value of ~118 million tons per year. They correctly point out, as did “Anonymous”, that if you add the Atchafalaya, the total extracted for their simulation would be only 25% of all that is available.

        Now, here is where it gets murky. Another really important number is the percent of total load that is allowed to go into the diversions, and is then trapped on the delta surface, rather than transported through. Kim et al. and Blum and Roberts use similar values, 35-40% of the total that is made available, based on studies of deltas elsewhere in the world.

        To sum up……Kim et al. predict:

        A. with 45% of the total lower Mississippi load of 118 MT/yr, and trapping 35-40% of that on the delta surface, they can build 700-1200 km2 by 2010.

        B. with 100% of the sediment from the Mississippi and Atchafalaya, and trapping 35-40% on the delta surface, the total area produced by the modeled diversions would be 2740 km2 by 2010. This included inflating the surface by 20% volume from organic matter.

        Back to Blum and Roberts……………..

        Blum and Roberts estimated that between 10,000 and 13,500 km2 (3860-5200 sq. miles) would be submerged by 2100 if there was no action taken. The 18-24 billion tons….thats the amount needed to sustain the whole thing in a state similar to today by filling in the space created by sinking of the surface and sea-level rise.

        Blum and Roberts assumed that 100% would be diverted (16.4 billion tons total) but that only 40% would be trapped. Hence, the total that would go into delta construction is ……..40% x 16.4 billion tons = ~6.4 billion tons…….. For this reason, Blum and Roberts predicted that there was only 1/3rd to 1/4th of the sediment needed to sustain the delta as it exists today.

        This equates to roughly enough sediment to sustain 3350 km2………take 1/3rd of 10,000 km2 =3300 km2 or 1/4th of 13,500 km2 = 3375 then round off a bit……

        And now compare………………

        Compare the estimate of 3350 km2 to the Kim et al value of 2740 km2, and there is a 20% difference, but the Blum and Roberts estimates would be more optimistic, not less, as has been implied.

        This difference is because of the way the 2 studies treat spatial variation in subsidence……..the Kim et al. study looks at a small area, so it does not matter, whereas Blum and Roberts look at the entire delta, and much of it will be subsiding at a lower rate, so more land can be built……. If they run their model using lower values of subsidence, they build more land, and reach the same estimate as Blum and Roberts.

        Hope this helps………………

        • Mike-
          I responded to David’s correction of my error before I saw your helpful comments. By now we’ve probably exhausted the patience of most readers!
          I appreciate all the comments, however.

  11. Your first replier, Anonymous, got it right. There is no significant difference between the Blum and Roberts estimates, and the recent estimates from EOS, as long as the underlying estimates of future sea level rise and subsidence are the same. The EOS estimates were first published in an LSU report. They were cited by Blum and Roberts as an example of the scale of the proposed solutions relative to the scale of the problem.

    On the sediments going off the shelf edge, yes, you have missed something. In the Blum and Roberts model, and the recent EOS paper, that IS the sediment that would be diverted, from the channel to the delta plain, and not allowed to go to the mouth and off the shelf edge.

    Also, not sure where the LACoast reference got their numbers of 159 MT/yr. The Blum and Roberts paper uses USGS data, and the Kim et al. EOS paper uses the same or similar numbers. It is more like 105-135 MT/yr for the Mississippi below Old River, and an extra 65-70 MT/yr for the Atchafalaya.

  12. I think the water freely moves throughout the locations mentioned.

    The Hydrologic Cycle.

    No question man has modified that transfer both in time and space.

  13. Hopefully Len won’t come too unglued for me questioning another item, but I think the math in Figure 1 needs a recheck.

    Believe a discrepancy exists between the caption and the two calendar years given.

    Sedimentation work is another one of those interesting arenas where one is never right; but one is never wrong either……. guess its now called ‘contrasting results….????’

    All the wrong answers carried out to six decimal places using a computer model…..

    As with power generation using water, is the land ‘really lost (consumed)’ or is it just moved around to a new location?????

    • HH-
      I certainly have my share of faults but at least my parts are still glued together. You’re absolutely right, and Figure 1 is being changed to describe 17 decades, not 7! Thank you.

      As for the global water cycle, of course it will continue with or without the 9 billion thirsty human beings projected to be alive in fifty years and its overall quantity will not change. What will shrink, however, is the current global ratio of one “teaspoonful” of clean freshwater to the “gallon” of saline and otherwise non-potable water.

      The huge volume of water used for thermoelectric power generation is “borrowed” from surface and ground water, depleting its availability for other local uses. When it returns to the atmosphere as water vapor it ends up someplace else, such as rain over open oceans.

      Freshwater is such an effective solvent that it is incredibly susceptible to contamination. Its use for irrigation, sewage disposal, industrial process water and thousands of other uses results in the net conversion of potable water to a dilute soup of chemicals that is unsafe to drink without expensive treatment.

      Global temperature increase that is rapidly destroying the mountaintop glacial storage and dependable slow release of freshwater threatens the deltas of many of the world’s major rivers, including the Ganges-Brahmaputra., so the growing millions of Bangladesh residents will experience extreme drought and flooding events.

      Oh well, not to worry.

  14. When 45% of the Mississippi is diverted, the model described in the EOS article predicts 700-1200 square kilometers of new land in 100 years. Kim et al. extend their calculations to include 1) all sediment in the lower Mississippi and Atchafalaya rivers (~210 megatons/yr) and 2) 20% organic inflation of deposit height. Under the extended conditions the model predicts 2740 square kilometers of land by 2100 utilizing the “base case” scenario for subsidence (5 mm/yr) and sea level rise (2 mm/yr). The predicted land building if all the Mississippi and Atchafalaya sediments are used (2740 square kilometers) represents ~1/4 of land loss estimated by Blum and Roberts (2009), and almost 1/2 of the land loss estimated by Barras et al. (2003).

Leave a Reply