| | LSU AgCenter researchers conducted a large-scale, multi-year study to assess the effects of dredge sediments on marsh recovery. Note the 40-acre study site south of Leeville. This satellite image is from 2002 and available on the U.S. Geological Survey Web site. |
| | | Figure 1. 1998 pre-dieback and pre-treatment aerial photograph showing sediment treatment, control and reference marsh sites.Both the treatment and control boundaries are overlayed for reference. Figure 2. Same aerial photograph as shown in Figure 1 but with the habitat layers highlighted for contrast. |
| | | Figure 3. 2001 post-dieback and pre-treatment aerial photograph showing three habitat signatures, remaining vegetation,dieback and water. Figure 4. Same aerial photograph as shown in Figure 3 but with habitat layers highlighted for contrast. |
| | | Figure 5. This visual is the 1998 image with the pre-sediment treatment locations of tidal channels outlined in red. Figure 6. This is a post-sediment treatment image (spring 2003) with the 1998 channel overlay. Note that virtually all of the original site hydrology has been changed and for the most part has been replaced with isolated discrete shallow water bodies. |
| | | Figure 7. This visual is from the fall of 2004 – 24 months post-sediment treatment. It shows both the sediment treatment and the control sites with the control boundary overlay, but without the habitat attribute layer. Figure 8. This is the same area, as at left, for both the treatment and control sites but with the habitat attribute layers highlighted for comparison. |
| | | |
Michael D. Materne, Irving A. Mendelssohn, Angela Schrift, Ashley Wilson and Sheila Rohwer
Wetland deterioration is a significant environmental problem in coastal Louisiana. Although natural and human induced factors have both been cited as causing wetland loss, many of these effects are mediated through one common agent: sediment availability. Therefore, a sound approach for reducing wetland loss and restoring deteriorated wetlands is the addition of sediment to increase marsh elevations to a level that will support wetland plants.
Although a number of innovative ideas—spray dredging, sediment transport through abandoned oil and gas pipelines, beneficial-use dredging—appear promising, a major limitation to the use of dredged material is the limited understanding of how different depths of added sediment affect the structure and function of wetlands. Too little sediment may have no beneficial effect while too much sediment may detrimentally modify the vegetative dynamics and ecosystem processes essential for the maintenance and self-regulation of these systems. Thus, there is a need to integrate additional biotic and physical elements into dredgesediment engineering to maximize the successful and beneficial use of dredge material for both large-scale marsh creation and small-scale marsh enhancement.
Because of the high cost associated with hydraulic dredging, this study was coordinated with the Louisiana Department of Natural Resources (LDNR) to provide a cost-efficient opportunity to study marsh nourishment on a landscapescale not normally available through traditional research funding. Collaborating with LDNR engineers, researchers from the School of Plant, Environmental & Soil Sciences and the Department of Oceanography & Coastal Studies were able to conduct a large-scale, manipulative and multi-year study to assess the effects of dredge sediments on marsh recovery following a catastrophic event.
Four primary objectives of this study were to:
- Assess spatial changes across treatments and controls throughout the life of the study.
- Construct a hydrologic model to assay the distribution, frequency and duration of flooding effecting treatments and controls.
- Assay physico-chemical factors controlling natural vegetative recruitment and recovery.
- Determine the effects of artificial plantings and supplemental nutrients on vegetative recovery, recruitment, succession, and above-ground productivity.
This study was conducted on about 40 acres of a large marsh die-back (about 110,000 acres) that occurred in Louisiana’s saline marshes in 2000 and 2001. The project design consisted of constructing five 6-acre contained cells that received hydraulic sediment pumped over ambient marsh in depths varying from 5 inches to 20 inches.
The project design also included two controls, which were four-acre unconfined cells that did not receive sediment, and two healthy reference (ambient) marshes adjacent to the project site. Sediments were hydraulically dredged material from Bayou Lafourche in the form of a sediment-slurry at a ratio of about 25 percent solids to 75 percent water-by-volume. The sediments were pumped into the marsh in the fall of 2002.
Soil, Plant Measurements
Approximately 210 linear sample points were established across the sediment treatment, control and reference cells. A number of vegetative responses—such as species recruitment, density and cover—were measured in the fall of 2003, in the spring of 2004, and again in the fall of 2004. Concurrent with the vegetative recruitment measurements, several physicochemical variables—such as reduction-oxidation potential, bulk density, organic matter, electrical conductivity, pH and soil particle distribution—were measured during each sampling period. In addition, extractable ammonium nitrogen, nitrate nitrogen, phosphorus and exchangeable calcium, magnesium, potassium, sodium, iron, manganese, copper and zinc were measured.
Within the sediment treatment, control and reference cells, we established 430 plots of six plant species treatments. Species included Avicennia germinans (black mangrove), Distichlis spicata (saltgrass), Juncus roemerianus (black needlerush), Spartina alterniflora (smooth cordgrass), Spartina patens (marshhay cordgrass) and unplanted. Half the vegetative plots received supplemental nutrients, and the other half received no additional nutrients. Response variables included survival, density, spread and several above-ground productivity measurements.
Spatial Change Assessment
A geographical information system (GIS) project was created by integrating a series of time-period base maps, using high resolution geo-referenced aerial photographs from which a geospatial dataset could be developed and subsequent change could be accurately measured across the life of the study. Six base maps were used to develop the geospatial indices. These included aerial photographs taken by the U.S. Geological Survey (USGS) in 1998 and 2001. The remaining four base maps were flown under contract for this study in April 2003, December 2003, December 2004, and December 2005.
We established common control polygons placed over each base map to restrict attribute delineations to a fixed constant and to provide a summary check to ensure accuracy across multiple base maps. Summary tables were indexed by attributes, and area measurements were created for each map series, compiled and used for comparison analyses.
Hydrology
An additional objective of this study was to assess the hydrologic-elevation effects on vegetative recovery resulting from sediment enhancement. To accomplish this we installed a YSI model 600LS sonde, equipped with multi-sensors and a vented level system, to record changes in both water level and salinity values. We measured water temperature, specific conductance, salinity and water levels at 30-minute intervals. The sonde logged approximately 21,000 entries continuously over a 15-month sampling period, including through both hurricanes Katrina (Aug. 29, 2005) and Rita (Sept. 24, 2005).
For quality control and quality assurance checks, a set of real-time measurements was taken at the end of each quarterly redeployment and checked against applicable standards before leaving the sonde unattended. In addition, a water sample was taken and analyzed under laboratory conditions as an additional check. Several statistical hydrograph models were constructed to assess flooding distribution, frequency, duration and depth across key elevations as well as to determine possible correlations between flooding, salinity and elevation.
Study Results: Spatial Change Assessment
Before the dieback event in 2000, we found a robust marsh with emergent vegetative cover within both the sediment treatment and control areas of the study sites. See Figures 1 and 2. Vegetation within the sediment treatment boundaries was the dominant feature with 91 percent cover followed by permanent open water at 9 percent. The control areas also contained vegetation- to-water in approximately the same proportion as that of the treatment area, with 98 percent vegetation and 2 percent water.
By November 2001 (post-dieback and pre-sediment treatment), we found extensive areas of vegetative loss. See Figure 3. Using the same boundaries and delineation methods used with the 1998 base map, we used the dark gray signature of the reduced plant stubble and exposed soils found on the 2001 maps to delineate habitat loss resulting from brown marsh effect within the study area. In addition to healthy emergent marsh and open water bodies, we added a third attribute to the 2001 dataset to account for impacted (dead) marsh as affected by the brown marsh events.
We found a significantly large amount of marsh loss in both the pre-sediment treatment and the control sites of the study area. Marsh loss within the study area followed the same uniform pattern seen throughout the Barataria-Terrebonne Basin. That is, the greatest mortality occurring in the marsh interior with lesser degrees of mortality along shorelines and tidal creek banks suggested a correlation with tidal flushing.
Comparing 2001, 2003
We compared habitat change from 2001 (brown marsh) to spring 2003 (four months post-sediment treatment) across three habitat variables — vegetative cover, impacted soils and permanent open water. Four months after treating the area with additional sediments, there was little statistical change across the sediment cells. By spring of 2003, there was a 10 percent reduction in vegetative cover from that found in 2001. Impacted acres remained static with less than a 1 percent increase, and open water increased nearly 18 percent.
Although there were relatively small statistical changes within test variables, there were significant physical changes in habitat distribution, particularly with vegetation and water bodies. For example, in comparing the location of healthy vegetation between 2001 and spring 2003, we noted that what remaining relatively healthy marsh can be seen in the 2001 aerials as a pinkish-red signature in Figure 3 and as the dark green overlay in Figure 4 is clustered around tidal creeks and contained in fairly discrete stands of vegetation.
In spring 2003, however, all of the larger stands have disappeared, and vegetation reappears as small, randomly distributed clumps not particularly associated with any apparent surface features such as water or pre-treatment standing crop. It would have been reasonable to assume that post-treatment vegetative recovery would have been concentrated in areas where pre-treatment vegetative communities existed. Historically, above- and below-ground standing crops have proven to be rich sources for accelerating vegetative recovery in marshes that have been heavily impacted or in newly created marsh. Figure 6 shows the scattered (green overlay) area of plant recovery four months post-sediment treatment. We observed in this study, however, a complete redistribution of plant materials with no apparent correlation to pre-treatment standing crop.
Water distribution changes paralleled that of vegetation, but the changes resulted from different mechanisms. For comparison, we delineated the location of tidal channels (red outline) on the pre-treatment 1998 imagery (Figure 5) and overlaid these channels on the post-treatment spring 2003 image (Figure 6). For contrast, we delineated and colored existing water (dark blue) and vegetation (dark green) on the 2003 image. Note that only in the immediate area, where the retention levees were intentionally breached (marked by white circles in Figure 6), was there some minimal return to the original watershed pattern. For the most part, water bodies within the treatment cells did not return to their original patterns. Rather, they re-established primarily along levee excavation lines.
Plant Cover Increases
For the purposes of this study, we defined net marsh loss as the combined losses in both emergent marsh and increases in water bodies — that is, emergent marsh that remains unvegetated (exposed soils) and an increase in water area from its original baseline value.
We found that at the end of the study period, the treatment sites had recovered to 94 percent of their pre-dieback marsh conditions with a net loss of 6 percent, while the control sites recovered to 87 percent of their pre-dieback conditions with a net loss of 13 percent.
We found no net losses within the ambient (reference) marshes that were not impacted by the 2000 dieback. Compare Figures 3 and 4 (post-dieback, pre-treatment) with Figures 7 and 8 (24 months post-treatment) for visual comparison of habitat change.
Study Results: Hydrology
To assess water and salinity effects for the study area, we constructed a number of hydrographs to determine flooding frequency, distribution, duration and depth across selected elevations, as well as to determine possible correlations between salinity, flooding and rainfall. Over the 15-month study period, we found tidal cycling to follow a diurnal (one high and one low tide per 24.8 hours, or one lunar day) cycle, modified by overriding weather effects such as rainfall and storm events. In addition, we found the study sites to be a strong saline marsh (average salinity 21 parts per thousand) and water levels fluctuating between minus 6.9 inches below and 19.8 inches above ambient marsh. The average water level for the 15-month period was minus 0.13 inches, or just slightly below the ambient marsh surface.
In addition, we found both water level and salinity varied seasonally with the highest water levels during the summer months and highest salinity levels during the fall months. We also found that there was no correlation between tide fluctuation and salinity but a moderately strong negative relationship between rainfall and salinity.
Percent Time Flooding (Duration)
We defined flooding duration as the cumulative amount of time (expressed as percent of total) that standing water remained at or above a specific elevation. Because each recorded occurrence is equivalent to a 30-minute duration, the cumulative sum of each occurrence at or above a specific elevation would equal the total time (or percent) that the marsh was flooded.
At ambient marsh (zero elevation), we found that there were almost equal periods of surface flooding (49 percent) and drying (51 percent). This nearly one-to-one ratio of wetting and drying is optimal for plant growth by providing a balanced oxidation-to-flooding regime, critical to intertidal plant species such as Spartina alterniflora.
Within the control site elevations, the ratio of flooding (88 percent) to drying (12 percent) was disproportionately skewed toward wet soils. Under these flooding conditions, there is a high probability of significantly reduced soils, resulting in lower plant survival and reduced plant productivity. Plant survival within the control cells, across all species and nutrient levels, was zero. When we compared the presence or absence of vegetation along a continuous elevational profile within control site A, we found that the treatment sites would not support vegetation when the marsh surface had eroded to greater than 1 inch below the ambient marsh elevation.
Physico-chemical, Plant Response
One of the primary factors influencing plant distribution and function in Louisiana intertidal saline marshes is elevation. In low intertidal marshes, soils are generally characterized as reduced, and Spartina alterniflora is the dominant plant species in part because of its superior oxygen transport mechanisms and high sulfide tolerance. As waterlogging (reduction) stress decreases with increased elevation, competitive interactions cause the replacement of S. alterniflora with other species.
Within the lower sediment elevation treatments (5 inches to 7 inches above ambient marsh), there was rapid S. alterniflora recruitment with densities similar to that found in the healthy reference marsh sites. With increased elevation (8 inches to 10 inches), we found reduced S. alterniflora recruitment, but a significant increase in species richness. Likewise, within the higher elevational treatments (11 inches to 14 inches), there was minimal S. alterniflora recruitment, and the greatest increase of species richness compared to the other treatment cells, controls or the reference sites.
At elevations greater than 18 inches above ambient marsh, however, there was little to no vegetative recruitment of any species, suggesting an upper threshold of hydrologic-soil-plant interaction. Within the control cells, recruitment and species diversity was marginal in areas that received no supplemental sediment. In the absence of vegetative cover, significant portions of the control cells collapsed and reverted to open water during the course of the study.
Frequency and duration of flooding and reduction-oxidation (redox) potential have been found to correlate with elevation in tidal salt marshes. Similarly, within the study area we found redox potential and elevation to be highly related, and we identified an inverse relationship between percent time flooded and elevation at our study site. The markedly reduced redox potentials seen in the control cells can be attributed to their low elevation and increased flooding frequency and duration.
Flooding duration within the control cells was 88 percent compared to 49 percent within the reference marsh. When soils are flooded, oxygen depletion is rapid because of its slow rate of diffusion and consumption by facultative and anaerobic microorganisms during respiration.
Sediment Decreases Sulfide
These microorganisms use oxygen as a terminal electron acceptor converting oxidized compounds into their reduced states, potentially forming toxins, such as hydrogen sulfide. High sulfide levels have been found to decrease productivity and are, therefore, detrimental to the plant growth. We found soil sulfide concentrations to sharply decrease with the addition of sediment, and treatment cells receiving sediments were found to have no measurable concentrations of interstitial sulfide.
In contrast, we found high sulfide concentrations and low root oxygen concentrations in the control cells that would have inhibited nitrogen uptake and contributed to high average ammonium nitrogen concentrations, thus causing a reduction in plant growth.
In addition, the concentration of salts such as magnesium, calcium, sodium and potassium throughout our study area was within normal ranges and in many instances not significantly different among or between treatment levels and marsh types. These results imply that salt concentrations were not a factor limiting plant colonization at any of the sediment elevations within the study site.
High bulk density (a measure of mineral content) has been shown to increase plant recovery and productivity. Mineral matter can improve marsh vigor by increasing nutrient availability and by decreasing toxicity via providing metals (iron and manganese) that precipitate with sulfide.
Bulk density was significantly higher in sediment-subsidized areas, and specific treatment levels had marked increases of iron, copper, manganese, zinc, calcium and phosphorus compared to reference marshes. Plant survival across all treatments ranged from excellent to extremely poor within the treatment and control cells. There were no surviving plants at the control sites, indicating that sediment addition was the critical factor in species survival.
When assessing nutrients and elevation as main treatment effects, we found no significant difference across any plant species treatments. For example, survival within the fertilized plantings averaged 72.7 percent compared to 71.6 percent in the unfertilized plantings.
Elevation Affects Species
When we combined elevation and species as treatments, we found that species segregated along elevation. For example, S. alterniflora, which is typically an intertidal species, performed better at lower elevations than at higher elevations while D. spicata, which prefers a less saturated soil, increased with increases in elevation. Avicennia germinans was the only species that performed equally well across all elevations. The control sites, which were considerably lower (about minus 6 inches below ambient marsh) had total mortality across all plants species treatment. It should be noted that S. alterniflora (the pre die-back dominant species) reached vegetative equivalency to that of normal marsh in less than two complete growing seasons across all treated cells, with the exception of the untreated control cells.
Wetland loss is indeed acute in Louisiana. The effects of increasing demands on petroleum exploration, global climate changes, worldwide rise in sea levels and human pressures for coastal development will remain critical issues well into the future. Research completed from this study provides a better understanding of the hydrologic-soil-plant relationships and has shown that the impacts of a large-scale, severe disturbance in a subsiding salt marsh can be minimized through the addition of sediment slurries.
When combined with other coastal restoration research, the information from this study will provide coastal wetland project planners, designers and builders additional management strategies, which will better incorporate vegetative diversity and productivity into beneficial use sediment engineering.
Michael D. Materne, Instructor, School of Plant, Environmental & Soil Sciences; Irving A Mendelssohn, Professor, and Angela Schrift, former Graduate Student, Wetland Biogeochemistry Institute and Department of Oceanography & Coastal Sciences, LSU, Baton Rouge, La.; Ashley Wilson and Sheila Rohwer, both Research Associates, School of Plant, Environmental & Soil Sciences, LSU AgCenter, Baton Rouge, La.
(This article was published in the spring 2007 issue of Louisiana Agriculture.) |