• • COMPARISON OF ECOSYSTEM STRUCTURE AND FUNCTION OF CREATED AND NATURAL SEAGRASS HABITATS IN LAGUNA MADRE, TEXAS • Paul A. Montagna, Principal ·investigator Cooperative Agreement No. X-00658801-0 Technical Report Number TR/93-007 • • • • • • • MARINE SCIENCE LIBRAR~ The University ofTexas at Austm 750 Cbannelview Drive Port Aransas , TX 78373 U.S.A. • • FINAL REPORT • COMPARISON OF ECOSYSTEM STRUCTURE AND FUNCTION . • OF CREATED AND NATURAL SEAGRASS HABITATS IN LAGUNA MADRE, TEXAS • by • Paul A. Montagna, Principal Investigator from • University of Texas at Austin Marine Science Institute P.O. Box 1267 Port Aransas, Texas 78373 • to Dr. James H. Ratterree, Project Officer • Water Management Division (6W-QM) U.S. Environmental Protection Agency, Region 6 1445 Ross Avenue Dallas, Texas 75202-2733 • Cooperative Agreement No. X-00658801-0 The University of Texas Marine Science ljistitute Technical Report Number • TR/93-007 November 3, 1993 • • COMPARISON OF ECOSYSTEM STRUCTURE AND FUNCTION OF CREATED • AND NATURAL SEAGRASS HABITATS IN LAGUNA MADRE, TEXAS TABLE OF CONTENTS • LIST OF FIGURES II LIST OF TABLES iii • ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 INTRODUCTION 2 • METHODS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Study Design ........... . ..... . ................. .. .... ·; . . 5 Study Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Study Site Descriptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .·. 6• Hydrographic Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Geological Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Chemical Flux Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Biological Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9• Statistical Analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 RESULTS ............................. , . . . . . . . . . . . . . . . . . . . . 12 Synoptic Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 • Temporal Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 DISCUSSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 • ACKNOWLEDGEMENTS 46 I REFERENCES 47 • • • LIST OF FIGURES • • Figure 1. Upper Laguna Madre. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Figure 2. Northern part of Upper Laguna Madre. . . . . . . . . . . . . . . . . . . . . . . . 4 Figure 3. Sediment composition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Figure 4A. Sediment eH profiles. Vegetated stations. . . . . . . . . . . . . . . . . . . . 20 • Figure 48. Sediment eH profiles. Unvegetated stations. . . . . . . . . . . . . . . . . . 21 Figure 5. Sediment organic matter at two depths. . . . . . . . . . . . . . . . . . . . . . 22 Figure 6. Sediment organic matter components. . . . . . . . . . . . . . . . . . . . . . . 23 Figure 7. Oxygen flux. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 Figure 8. Total dissolved inorganic nitrogen flux. . . . . . . . . . . . . . . . . . . . . . . 25 Figure 9. Ammonia flux. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 Figure 10. Nitrite flux. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Figure 11. Nitrate flux. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 • Figure 12. Phosphate flux. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Figure 13. Silicate flux. . ......................................~ . 30 Figure 14. Macrofauna abundance at two sediment depths. . . . . . . .. . . . . . . . 31 Figure 15. Macrofauna biomass at two sediment depths. . .............. . 32 • Figure·16. Macrofauna taxa abundance. 33 Figure 17. Macrofauna taxa biomass. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 Figure 18. Macrofauna species principal factor analysis. . . . . . . . . . . . . . . . . . 35 Figure 19. Macrofauna species diversity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36• Figure 20. Macrofauna species evenness. . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 Figure 21. Relationship be~een species diversity and evenness. . . . . . . . . . . 38 Figure 22. Sediment organic matter at two sediment depths over one year. . . . 39 Figure 23. Sediment organic matter components over one year. . . . . . . . . . . . 40• Figure 24. Macrofauna abundance at two sediment depths over one year. . . . . 41 Figure 25. Macrofauna biomass at two sediment depths over one year. . . . . . . 42 • • • ii • • • Table 1. Table 2. Table 3. Table 4. • Table 5. Table 6. Table 7. Table 8. Table 9. LIST OF TABLES Sampling locations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 Hydrographic measurements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Sediment grain size in Laguna Madre. 53 Eh profiles in sediment cores. 54 Oxygen measurements in sample incubations. . . . . . . . . . . . . . . . . . 55 Nutrient measurements in sample incubations. . . . . . . . . . . . . . . . . . 59 Vertical distribution of macrofauna in April 1992. . . . . . . . . . . . . . . . 63 Species distributions in April 1992. . . . . . . . . . . . . . . . . . . . . . . . . . 66 Laguna Madre Diversity and Evenness. . . . . . . . . . . . . . . . . . . . . . . 71 Table 10. Temporal changes in sediment organic matter. • • • • • • • iii . . . . . . . . . . . . . . . . 72 • COMPARISON OF ECOSYSTEM STRUCTURE AND FUNCTION OF CREATED • AND NATURAL SEAGRASS HABITATS IN LAGUNA MADRE, TEXAS ABSTRACT • • There is increasing demand to mitigate the loss of submerged wetland habitats. This project is designed to identify the criteria for a successful mitigation project, and the time for a created seagrass bed to become a functional habitat. Two approaches are taken. The first is a synoptic study of mitigated sites of different ages, the second • • is monitoring of a recent mitigation site for one year. Ecosystem structure and function is assessed by measuring select variables. Community metabolism and nutrient regeneration are key variables, which indicate the functioning of an ecosystem. Benthic community structure is a key variable that indicates the habitat utilization of an ecosystem. The mitigation sites are compared to three natural reference sites. Above­ground, the mitigation sites resembled natural sites in terms of biogeochemical f~.mction, but there were large differences below-ground. The mitigation sites lack sufficient organic material in the sediment for the environment to be fully functional. Benthic community structure at the mitigation sites resembled disturbed environments with high number, diversity, and low evenness. There was also a discernible trend among sites of different ages, that suggest it may take longer than 14-17 years to fully recover. Since this is such a long time, monitoring for one year did not reveal these differences . • Future projects to transplant seagrasses for mitigation should consider adding organic matter to the soil to speed the time it takes for the habitat to become fully functional. • • •; • 1 • INTRODUCTION • • Seagrass habitats are important to desirable fish and wildlife species (Kikuchi, 1974). Yet, numerous seagrass habitats have been damaged or destroyed by discharges, dredging and marine construction in our nation's bays and estuaries. There have been many projects to mitigate these adverse impacts on fish and wildlife. The • general · goal of mitigation programs is to replace habitat or repair damage. National Marine Fisheries Service recommends that mitigation projects should attempt to reestablish wetland fishery habitats and their ecological function (Thayer et al., 1986). Mitigation projects generally include the restoration or creation of new seagrass habitats, but monitoring or evaluation of the success of these projects is rarely done. • When it is performed it is usually limited to describing the success of the plantings. For example, in south Texas estuaries four out of seven mitigation projects planted between 1978 and 1983 were judged successful (Cobb, 1987). Success was determined by comparing percent cover in the mitigated area versus a control area. Much less is not known on whether these mitigated habitats are functioning like natural s~agrass habitats. Biological interactions between plants, animals, and microbes have a profound • effect on the success of any habitat creation project. After initial construction or planting, there is a succession of events leading to the climax, mature seagrass community. This process includes colonization of the unvegetated or transplanted area by microbes, epiphytes, and benthic invertebrates. The microbial community is• important in maintaining the balance of available nutrients, which are necessary for plant growth. Invertebrate bioturbation plays an important role in irrigating sediments with water and oxygen, which can enhance nutrient cycling rates. Finally, a luxuriantly vegetated benthos can provide the habitat for a variety of vertebrate and invertebrate • species. All these processes must occur before the mitigated environments become a functioning habitat in the sense of an ecosystem. The objective of this study is to compare benthic metabolism, nutrient regeneration, and habitat utilization of created seagrass habitats of different ages with • natural habitats. The goal of collecting this data is to determine how, and when created • habitats become functioning ecosystems like natural systems. This information is neces~ary to define measures of success, and delineate how long it takes a planted system to provide the ecological functions that are provided by naturally occurring seagrass systems. This information can alsb be used to. develop new criteria or methods for projects to create, enhance or restore seagrass habitats . • 2 • • • • • • Figure 1. Upper Laguna Madre. Natural reference sites are in Baffin Bay (6), and the Laguna Madre (189). Mitigation sites are between the shorelines of the cities of Flour Bluff and Padre Isles . • • • • • 3 • • • Is I es • • _­ • • . I 20 • • e • • • • • • • Figure 2. Northern part of Upper Laguna Madre. Location of mitigation sites. Channels are shown in dashed lines . • • • 4 • • • • • • • • • • • . . . . . . . : :: .. • METHODS • Study Design • Two studies were performed. One study was a synoptic sampling of 13 stations to compare community structure and rates of biogeochemical processes at natural, old • and recent mitigation sites. Three stations were naturally vegetated sites, nine stations were in mitigation sites and one was in a muddy bottom of an open bay. Two of the mitigation sites were constructed in the mid-1970's and are about 14-17 years old. These are called "old sites". Three of the mitigation sites were constructed between • 1990 and 1991 and were 1-2 years old when s~mpled. These are called "new sites". It is reasonable to assume that natural sites are much greater than 20 years old, so the natural sites represent the oldest sites. An important feature of the study design is that we are replicating sites,· that is replicating at the treatment level to avoid • pseudoreplication. The 13 stations used for the synoptic study were sampled in April 1992. The second study was performed to monitor seasonal variability in community structure at a natural and mitigation site. Four stations, two natural and two mitigation, were sampled quarterly throughout a one-year period. Study Area • Ten study sites were chosen in the Upper Laguna Madre and Baffin Bay (Table 1). Two of the sites have been visited since 1989 as part of a long-term research project to determine the importance of seagrass beds in maintaining a productive finfishery (Figure 1). These sites are 189 in the southern upper Laguna Madre and 6• in Baffin Bay (Table 1). Eight of the sites were located in a small area in the northern Upper Laguna Madre between the Flour Bluff and Padre Isles shorelines (Figure 2). In most cases there is only one station per site. At three sites, there are two paired station locations. One station is located in the grass bed, and one station is• adjacent in a bare patch. These paired stations are located in sites 189, TS and GI. The suffix (-G) for the grass and (-S) for sand patch is used to name each site: 189G, 189S;··'TSG, TSS, GIG, and GIS. Only station 6, which was in mud, does not have a suffix added to the station name. ·· ~ • All stations were sampled during the synoptic study in April 1992. Four stations (189G, 189S, TSG and TSS), one at a natural site (189) and the other at a mitigated site (TS) were sampled in each of the four seasons during the temporal study. I' • 5 • Study Site Descriptions • Gulf Isles Limited (GI_), project #9009(08) is located east of lntracoastal Waterway Marker 49. The project scraped down a spoil island and created an area of submerged habitat approximately 320 m x 168 m with six circulation channels in April, • 1991. Seagrass planting was not required. Natural colonization by Ruppia maritima, • Ha/odule wrightii and Ha/ophila engelmannii appears to have been successful. Two stations were sampled in the southern end of the excavation site at a depth of 0.4 m. One station was in a mixed bed of H. enge/mannii and R. maritima (GIG) and the other was an adjacent bare sand patch (GIS). The sediment was firm in both areas composed of approximately 90% sand, 5% rubble, 2% silt and 3% clay. • Padre Isles Natural Site #1 (Pl 1 G) was lo'cated in an open area east of the Gulf Isles site and west of the Padre Isles development. This site is protected from high wave action due to the surrounding land resulting in a low energy area. Most of this area is covered with a mixture of H. wrightii and R. maritima with few bare patches. Core samples were taken from a bed of H. wrightii at a depth of 0.5 m. The sE1diment was very soft and smelled of H2S when disturbed. The upper 3 cm of sediment was composed of 10% rubble, 55% sand, 10% silt, and 25% clay while from 3 to 10 cm • depth sand increased to 90%. Padre Isles Natural Site #2 (Pl2G) was located in the center of a seagrass flat east of the spoil islands adjacent to lntracoastal Waterway Marker 63 and west of Padre Island. The dominant seagrass at this site is H. wrightii. Samples were taken • at a depth of 0.75 m in a bed of H. wrightii. The sediment was firm compared to Pl1 with more rubble 14% and sand 74% and less silt 25% and clay 10%. The deeper sediment (3-10 cm) had higher sand content (89%). Transco scrape-down (TS_) project #18853 is located in state land tract 64 on • a spoil island east of lntracoastal Waterway Marker 55. Submerged habitat was developed by scraping down an existing spoil island, cutting three circulation channels and planting H. wrightii. Samples were taken from H. wrightii (TSG) and bare sand (TSS) in a water depth of 0.4 m. The sediment was very firm composed primarily of • sand in the grass (88%) and the bare patches (95%) . Transco pipeline (TPG) project #18853 was an attempt to establish seagrass, H. wrightii, on the bare shoulders of a pipeline extending from Padre Island in state land tract #17 4, and 64 under the lntracoastal Waterway near Marker 59, and through state tracts 48, 47, 25 and 134 to the mainland~ The site sampled was located east of the spoil islands adjacent to marker 59 near the area where the pipeline crossed the state tract boundary between state tracts 64 and 174. The water depth was 0.6 m and the dominant grass along this section of the pipeline was R. maritima. The sediment l' " 6 • • • • • • • • • • was firm composed of 95% sand. Central Power and Light Company (CPG) project #10444 is located on the west Laguna Madre shoreline adjacent to the CP&L mariculture ponds. The project resulted in the removal of dredged material covering submerged seagrasses and was described as being successful. The project site is in a small cove formed by a point of land to the north with the opening facing the southeast. The predominant southerly winds deposit dead seagrass along the shoreline and on the bottom. Ruppia maritima was the dominant seagrass and was sparse. The water depth was 0.55 m and the sediment was very soft. The upper 3 cm of sediment sampled was 9% rubble, 70% sand, 12% silt, and 9% clay and the 3-10 cm sediment layer was 10% rubble, 79% sand 1 % silt and 10% clay. Skyline Equipment, Inc. (SKG) project #12004 (03) is located on the west Laguna Madre shoreline just north and adjacent to the Central Power and Light project. The project created 0.14 ha (0.34 acre) of submergent habitat from uplands in 1978 . The site is located on a point and is exposed to high energy southeast and northerly winds resulting in minimal dead seagrass deposition. The bottom was cove~ed with approximately 25% Ruppia maritima, 25% Ha/odule wrightii and 50% bare sand. Core samples were taken in H. wrightii at a depth of 0.35 m. The sediment was composed mainly of firm sand (92%). Marker 189 (189_) is a natural reference site in an open grass flat to the west of lntracoastal Waterway Marker 189. This site is vegetated with Halodu/e wrightii with scattered bare patches and very little drift algae and dead seagrass debris. The water depth is 0.8 m. Samples were taken from the grass (189G) and an adjacent bare patch (189S). The sediment in th~ bare patch sampled was firm composed of 21 % rubble, 61% sand, 3% silt and 15% clay. The grass sediment was similar with 21% rubble, 50% sand, 4% silt and 19% clay. The amount of clay increased with depth (35%) in the sandy bare patches and the seagrass. Genesis Petroleum (GES) project #15844 is located between two dredge spoil islands east of lntracoastal Waterway Markers 67. Approximately 0.4 ha (0.9 acre) of submerged wetland was created from the emergent spoil island. The site is in a small cove which faces southeast into the prevailing wind. Dead seagrass and detritus collect along the shoreline and on the bottom. Although H. wrightii was planted following the scrape-down, no living seagrass was found at the site. The water depth was 0.9 m. The surface sediment was 63% ..sand and 31%_. clay. Below 3 cm the '\ . ­ sediment was 94% sand. Marker 6 (BB6) is a control site located approximately 180° off of Marker 6 at the mouth of Baffin Bay. This site is in the open bay in 2.2 m water depth without seagrass. The sediment is soft mud predominantly silt (15%) and clay (81 %). 7 • Hydrographic Measurements • Salinity, conductivity, temperature, pH, dissolved oxygen, and redox potential were measured at each station during each sampling trip with a multiparameter instrument (Hydrolab Surveyor II). The sonde unit was lowered to just beneath the • surface and to the bottom. The instruments allows us to collect a variety of water • quality parameters rapidly. The following parameters are read from the digital display unit (accuracy and units): temperature (± 0.15 °C), pH (± 0.1 units), dissolved oxygen (mg/I ± 0.2), specific conductivity (± 0.015 -1.5 mmhos/cm depending on range), redox potential (± 0.05 mV), depth (± 1 m), and salinity (ppt). Salinity is automatically corrected to 25°C. Suspended sediments are measured as turbidity in nephelometric turbidity units (NTU) with a Hach photometer. Turbidity can be converted to suspended sediment concentration by making a standard curve of turbidity versus dry weight of filtered• sediments. In most Texas bays there is a linear relationship between suspended sediment and turbidity (R2 =0.99): suspended sediment (mg-ml-1) =0.038xNTU + 0.085 (Montagna, 1989). • Geological Measurements Sediment grain size analysis was also performed. Sediment core samples were taken by diver and sectioned at depth intervals 0-3 cm and 3-10 cm. Analysis followed • standard geologic procedures (Folk, 1964; E. W. Behrens, personal communication) . • Percent contribution by weight was measured for four components: rubble (e.g. shell hash), sand, silt, and clay. · A 20 cm3 sediment sample was mixed with 50 ml of hydrogen peroxide and 75 ml of deionized water to digest organic material in the sample. The sample was wet sieved through a 62 ΅m mesh stainless steel screen using a vacuum pump and a Millipore Hydrosol SST filter holder to separate rubble and sand from silt and clay. After drying, the rubble and sand were separated on a 125 ΅m screen. The silt and clay fractions were measured using pipette analysis . • Chemical Flux Measurements Biogeochemical fluxes were measured in the same 6.7 cm diameter core tubes • that were used to sample macrofauna. Samplrs were taken _by hand to a depth of 1 O cm by divers. Three replicates were taken within a 2 m radius. The water level was brought to the top with added station water. After settling for about 1 O minutes the initial water subsample was taken. Then the cores were closed with rubber stoppers • 8 • that had an oxygen probe and a relief valve so that a tight seal could be o~tained. • Cores were incubated in the dark for two hours. Ice chest coolers were used as incubation chambers. The coolers had station seawater circulated through them, via a pump, to maintain the temperature as near to ambient conditions as possible. Three replicate cores were used to determine sediment metabolism and nutrient regeneration. • One station water sample was incubated as a control for oxygen metabolism, and two control samples were incubated for nutrient regeneration. The controls were used to represent changes in the overlying water that were not due to the presence of the sediment. • Oxygen concentration changes were measured every 15 min using pulsed • oxygen electrodes (Endeco, Inc., Marion, MA). ~hese electrodes are of a recent design in which the measurement of oxygen concentration is flow-insensitive (Langdon, 1984). The electrodes are connected to a Pulsed D.0. Sensor™ that controls the timing of the electrical pulses sent to each probe. Data is interpreted by the Pulsed 0.0. Sensor • and logged automatically on a portable computer. Oxygen changes per unit time were estimated using linear regression analysis. Water subsamples were taken from the overlying water in the cores after the two hour incubation period to measure changes in other chemical constituents. Dissolved • inorganic nitrogen (DIN) concentrations of ammonia, nitrate, and nitrite and phosphate and silicate were measured from the water subsamples using highly precise autoanalyzer techniques (Whitledge et al., 1986). Nutrient changes were estimated as the difference from initial and ending values. The mean of two replicates was used as the control value. The flux (FLUX) for both oxygen and nutrients is calculated a function of the chemical change (CHANGE) with ·respect to time minus a control value, and was adjusted for the area of sediment (FACTOR) covered by the core and the volume • (VOLUME) of water contained in the core: FLUXmmol·m-2·h-1 =VOLUME I x CHANGE mmol·l-1 ·core-1 ·h-1 x FACTORm-2/core (1) • Biological Measurements • Sediment was collected from the same 6. 7 cm diameter core tube, that was used ~ . to measure chemical flux. The macrofauna were sectioned at depth intervals of 0-3 cm and 3-10 cm (Montagna and Kalke, 1992). Samples were preserved with 5% buffered formalin, sieved on 0.5 mm mesh screens, sorted, identified, and counted . • 9 • Each macrofauna sample was also used to measure biomass. Individuals were • combined into higher taxa categories, i.e., Crustacea, Mollusca, Polychaeta, Ophiuroidea, and all other taxa were placed together in one remaining sample. • Samples were dried for 24 h at 55 °C, and weighed. Before drying, mollusks were placed in 1 N HCI for 1 min to 8 h to dissolve the carbonate shells, and washed with fresh water. Sediment organic matter was also measured from each core. The seagrass stems, roots, and detritus from each sample was collected on a 0.5 mm sieve, dried and weighed . • Statistical Analyses • Macrofauna diversity is calculated using Hill's diversity number one (N1) (Hill, 1973). It is a measure of the effective number of species in a sample, and indicates the number of abundant species. It is calculated as the exponentiated form of the Shannon diversity index: • N1 =eH' (2) As diversity decreases N 1 will tend toward 1. The Shannon index is the average uncertainty per species in an infinite community made up of species with known proportional abundances (Shannon and Weaver, 1949). The Shannon index is calculated by: • . l (3) Where ni is the number of individuals belonging to the ith of S species in the sample • and n is the total number of individuals in the sample. Richness is an index of the number of species present. The obvious richness index is simply the total number of all species found in a sample regardless of their abundances. Hill (1973) named this index NO. Another well known index of species • richness is the Margalef (1958) index (R1). R1 is based on the relationship between the number of species (S) and the total number of individuals (n) observed: • R1-S-1 ' (4) ln(n) ) -. Although common, this relationship presupposes that there is a functional relationship • 10 • • between S and n. This assumption may not be justified in all cases. Evenness is an index that expresses that all species in a sample are equally abundant. Evenness is a component of diversity. Two evenness indices, E1 and E5, have been calculated. E1 is probably the most common, it is the familiar J' of Pielou (1975). It expresses H' relative to the maximum value of H': • Et=_!!!_= ln(N1) (5) In(S) ln(NO) • E1 is sensitive to species richness. E5 is an index that is not sensitive to species richness. E5 is a modified Hill's ratio (Alatalo, 1981): • E5 (1/A.)-1 Nt-1 5 n(n.-1) where,A.= L ' ' (6) i-1 n(n-1) • .A is the Simpson (1949) diversity index. E5 approaches zero as a single species becomes more and more dominant. \ Statistical analyses to reveal differences among sampling periods, stations and sediment depths were performed using general linear model procedures (SAS, 1985). • Analyses were performed on chemical flux and species abundance, biomass and diversity measurements. Two-way analysis of variance (ANOVA) models were used where sampling dates and stations were the two main effects or where stations and sediment depth were the main effects. One-way ANOVA was used to compare stations • during the synoptic study of natural and mitigation sites in April 1992. Grthogonal linear • contrasts were used to test five a priori hypotheses about the structure and function of the habitats studied (Kirk, 1982). The first hypothesis is that there is a difference between the means of all vegetated and all nonvegetated stations. The second hypothesis is that among seagrass stations, there is a difference between the means -. of the natural and mitigation sites. The third hypothesis is that there is a linear or temporal difference among ages of seagrass bed habitats; the natural sites are considered the oldest, CPG and SKG are considered the same age and designated "old" mitigation sites; and TPG, GIG, and TSG dre considered -as "new" mitigation sites. The fourth hypothesis is that there are differences among the means of the old and new mitigation sites. The fifth hypothesis is that there is a difference between the TRANSCO scrapedown (TSG) and pipeline seagrass sites (TPG). Tukey multiple • 11 • • comparison procedures were used to find a posteriori differences among sample means (Kirk, 1982). The stations means are reported in a Tukey test, and those that are not different to the 0.05 level are joined by underlining. Multivariate ANOVA was used to test for treatment effects on species data. Factor analysis with rotated and unrotated factors was used to determine if communities were similar in different stations . • RESULTS Synoptic Experiment • • The stations were all hydrographically similar in April 1992 during the synoptic study (Table 2). Salinity and temperature averaged 24.6 ppt and 24.0 °C respectively at all stations. Dissolved oxygen and pH averaged 7.54 mg.i-1 and 8.97 respectively. There were some differences in oxygen concentration due to site differences and sampling at different times of the day. Baffin Bay was the only site with high turbidity. There was considerable difference in sediment composition at all sites (Table 3; Figure 3). Baffin Bay was the only site dominated by mud, having a high silt ~nd clay content. In the natural site of the southern part of the study area, station 189, sand • composed half of the content of sediments. The southern natural site, 189, was no more than 55% sand. All the northern stations, natural and mitigation, were composed of at least 73% sand. Within sites, bare patches had 5-10% higher sand content than vegetated sediments. The seagrass obviously promotes settling of fine particles, since • there was a higher amount of silt and clay at these stations. Eh decreased with sediment depth at all stations (Table 4; Figure 4). There were dramatic differences among sites in sediment Eh profiles. Vegetated sediments (Figure 4A) were always much more negative than bare-patch sediments within sites • (Figure 4B). There was a gradient of electronegativity from recent mitigation sites to older mitigation sties to natural sites. The two new sites (GI_, TS_, and TP) had almost no vertical differences in Eh. This indicates that there is a lack of reducing power in sediments of recent mitigation sites. • There was a considerable amount of seagrass-derived organic matter in all samples, except for the unvegetated sediments (linear contrast, P=0.0001, Figure 5). In the natural and old mitigation sites, most of this material was associated with the surface of the sediment (Figure 5). There w.as more material in natural sites (934 g.m-) than in mitigation sites (438 g.m-2) (line~r contrast, P='0.0001). There was also • 2 a significant difference with age of the mitigation site (linear contrast, P=0.0001). Old sites had 619 g.m-2, but new sites had only 317 g.m-2• New sites had proportionately _ ... lesser amounts of all components (Figure 6), but especially less below-ground material, • 12 e.g., roots and detritus. The new total amount of material at new mitigation sites was not significantly different from unvegetated sediments (Tukey test). The general trend was for higher amounts of organic material in natural and newer mitigation sites and higher amounts in seagrass stations (mean dry weight in g.m-2, station name, and Tukey test): 985 952 876 793 445 357 326 267 220 165 18 12 5 Pl1G Pl2G 189G SKG CPG TPG TSG GIG 189S GES GIS TSS 6 Oxygen measurements collected from the oxygen electrodes for calculating oxygen metabolism is given in Table 5. Mean oxygen flux was calculated using equation 1 and is presented in Figure 7. The average oxygen flux is negative indicating that the sediments were consuming oxygen in the dark. Seagrass bed samples had the greatest oxygen demand, -8.0 mmol 0 2 · m-2 · h-1 compared to -1.1 mmol 0 2 . m-2. h-1 in non-vegetated sediments, because of the high biomass of the seagrasses themselves (linear contrast, P=0.0001). Average flux (mmol 0 2 -m-2 .h-1) at natural stations was -10.4, old mitigation sites was -7 .2, and new sites was -6.5). There was a trend of higher oxygen consumption with age of the habitat (linear contrast, P=0.0001 ). The sand and mud stations were not significantly different from one another. The general trend was for higher amounts of oxygen consumption at seagrass stations, and less at mitigation sites (mean flux in mmol 0 2 . m-2. h-1, station name, and Tukey test): 0.1 -0.9 -1 .4 -1.7 -1.9 -4.8 -5.1 -5.4 -7.1 -7.7 -7.8 -9.7 -16.4 GIS GES 6 189S TSS CPG GIG TSG Pl2G TPG 189G SKG Pl1G Nutrient measurements for calculating nutrient regeneration is given in Table 6. Flux for all nitrogen components, DIN, phosphate and silicate were calculated. Total DIN flux was near zero at most stations (Figure 8). There was a great deal of sediment nitrogen uptake in the southern stations. However, variability was so great, that is it difficult to detect differences among stations (average flux in mmol DIN-m-2 -h-1, station name, and Tukey test): 13 • 2.0 ·1.2 1.1 0.2 -0.8 -0.3 -0.9 -1.0 -1.0 -1.6 -5.7 -16.7 -26.3 • GIG TSS Pl1G TSG TPG Pl2G CPG 189S SKG GES GIG 189G 6 Ammonia flux was the greatest constituent of DIN. Ammonia flux was similar at all stations (Figure 9, average flux in mmol NH4. m-2. h-1, station name, and Tukey test): • 2.3 2.3 2.2 0.9 0.1 0.07 -0.2 -0.3 -1.0 -1.1 -1.4 -16.8 -26.0 TSS GIG TSG Pl1G 189S TPG GIS Pl2G CPG SKG GES 189G 6 • Nitrite flux generally, was near zero, but on average there was efflux (0.084 mmol N02 .m-2 .h-1). The only stations with a large amount of nitrite regeneration were the mud and natural seagrass station in southern Laguna Madre (Figure 10). Because of the high value at the mud site (station 6), there was more nitrite regeneration in • • unvegetated stations (0.20 mmol N02 .m-2.h-1) than in vegetated sediments (0.012 mmol N02 .m-2 .h-1) (linear contrast, P=0.0096). Except for the high values at 6 and 189G, there were little differences among stations (average flux in mmol NQ2.m-2 .h-1, station name, and Tukey test): 1.12 0.64 0.09 0.03 0.03 0.02 -0.03 -0.04 -0.05 -0.06 -0.08 -0.28 -0.29 6 189G TSG Pl1G TPG GIS TSS 189S CPG GES GIG SKG Pl2G • Nitrate flux was also generally near zero, but on average there was uptake by sediments (-0.77 mmol N03 -m-2-h-1) (Figure 11). The only stations with a significant amount of nitrate uptake were generally unvegetated stations (-1.60 mmol N03 .m-2 .h-1), • which were different from vegetated stations (-0.26 mmol N03 •m-2. h-1} (linear contrast, P=0.0001 ). The only station with a large amount of nitrite flux was GIS (Figure 11, average flux in mmol NQ3.m-2 .h-1, station name, and Tukey test): 0.43 0.30 0.13 0.11 -0.16 -0.17 -0.21 -0.37 -0.52 -1.06 -1.08 -2.11 -5.33 SKG Pl2G CPG Pl1G GES TPG GIG 6 189G 189S TSS TSG GIS -.. • Phosphate flux was not significantly different at any of the 13 stations (P=0.3206, one-way ANOVA). The mean flux was -0.265 mmol P04 .m-2.h-1, and was not different from zero (Figure 12). On average silicate was generated by sediments (5.1 mmol • 14 • Si04 -m-2 -H1) (Figure 13). Silicate regeneration was higher in the natural seagrass sites • (16.0 mmol Si04 -m-2 -h-1) than in the mitigation seagrass stations (2.4 mmol Si04 -m-2 -h­ • 1) (linear contrast, P=0.0001). Silicate flux was high in seagrass bed stations (7.5 mmol Si04 -m-2 -h-1) and low in non-vegetated stations (1.1 mmol Si04 -m-2 -h-1) (linear contrast, P=0.0001). This trend was driven by large fluxes at two natural stations (Figure 13, average flux in mmol Si04 -m-2 -h-1, station name, and Tukey test): 34.4 15.5 6.9 4.7 3.7 3.3 1.9 0.8 0.7 -0.3 -1.4 -2.0 -2.5 189G Pl1G 6 GIG TSG TPG GES TSS CPG SKG GIS Pl2G 189S • Macrofaunal invertebrates were much more abundant in the top 3 cm of surface sediment (Table 7, Figure 14). There were on average 19,994 animals-m-2 in top 3 cm, and 3,831 animals-m-2 in the 3-10 cm depths. There were significant interactions • among sediment depths and stations (2-way ANOVA, P=0.001), so it is difficult to determine if mitigation affected the vertical distribution of organisms. The percent of organisms present in the top 3 cm of sediment was calculated and a 1-way ~NOVA indicated there were station differences (P=0.0069). Total percent abundance in • surface was higher (linear contrast, P=0.0006) at all the seagrass sites (85%) than at • the unvegetated sites (72%). The average biomass at all stations in the top 3 cm of sediment was 5.11 g-m-2, and 5.84 in the 3-10 cm section. Differences in vertical profiles among stations were found for biomass (Figure 15, 1-way ANOVA, P=0.0003). Again there was a higher percentage of the biomass found in vegetated sediments • (63%) than in unvegetated sediments (38%) (linear contrast, P=0.0004). Natural sites had a higher percentage of the biomass in surface sediments (75%) than mitigation sites (55%) (linear contrast, P=0.0159). There was also an increased percentage of biomass in surface sediments with age of the seagrass bed; the old mitigation sites had 66% of the biomass in the surface, and the new sites had 48% at the surface (linear contrast, P=0.0228). The following is a Tukey test of the percent of biomass in the surface sediment: • 87.5 82.1 80.2 73.0 61.1 58.6 55.9 49.2 45.0 39.1 26.7 24.7 11.3 Pl2G Pl1G 6 -CPG TPG SKG 189G GIS GIG TSG 189S GES TSS In general, vegetated sediments had higher total abundances to a depth of 10 cm (32,229-m-2) than unvegetated sediments (10,098-m-2) (linear contrast, P=0.0001) . • 15 • • Natural sites had higher abundances (40,781.m-2) than mitigated sites (27,097.m-2) (linear contrast, P=0.0005). Although there was no difference among old (28,932.m-2) • and new sites (25,874.m-2) (linear contrast, P=0.4824), the trend of natural to old to new is significant (linear contrast, P=0.0006). Total macrofaunal density was highest in the vegetated (natural sites of the northern part of the study area (Tukey test, average numberx10 3 .m-2 to a depth of 10 cm, and station name): 55.8 41.1 29.0 28.8 27.1 27.0 25.4 25.4 25.1 9.1 6.1 4.3 3.8 Pl2G Pl1G SKG CPG GES TSG GIG 189G TPG 189S GIS 6 TSS • The average infauna! biomass in the top 10 cm of sediment was different only • among vegetated and unvegetated sediments (Figure 15, linear contrast, P=0.0245) . Although infauna! biomass varied by an order of magnitude there were few stati.stically significant differences among the stations (one-way ANOVA) (average biomass ;;in g.m-2 to a depth of 10 cm, and station name): • 21.3 19.1 18.8 16.4 14.2 13.3 10.6 9.6 5.2 4.6 4.2 2.6 2.5 TSS GIG SKG TSG CPG 189G TPG Pl1G Pl2G 189S GIS GES 6 • Community structure, in terms of major taxa, was different among the stations sampled (Figure 16). There were large differences among vegetated and unvegetated I • sediments (MANOVA, P=0.0001). Natural and mitigation sites were also different (MANOVA, P=0.0003). There were significant differences with respect to age of the vegetated sites (MANOVA, P=0.0006). The differences among sites was driven by • changes in polychaete density, since they generally dominated the communities in all stations. Polychaetes generally dominated biomass also (Figure 17). Differences similar to abundance were found among vegetated and unvegetated sediments (MANOVA, P=0.0017), natural and mitigation sites (MANOVA, P=0.0053), and with respect.. to age of the vegetated sites (MANOVA, P=0.0146) . .,.·Community structure, in terms of species distributions was also different among the stations (Table 8). The most obvious factor.that is related to changes in community • structure is whether the station is vegetated o~ unvegetated ·(Figure 18). This factor, factor 1 in Figure 18, accounted for 53% of the variability in species distributions. The second factor, which accounts for 23% of the variability, seems to be related to the age of the mitigation site. All natural stations, the oldest mitigation stations (CPG and • 16 • • SKG), and the pipeline site (TPG) group together in the center and left side of the second factor axis. The newer sites (TSG and GIG) group together with unvegetated sites on the right side of the second factor axis. • Species diversity is highest in seagrass systems (Figure 19, Table 9). The average N1 diversity for seagrass beds was 10.4 species compared to a 6.3 species for unvegetated stations (linear contrast, P=0.0001 ). Species diversity is highest in the • recently disturbed environments. Natural sites had a lower average diversity (8.0) than mitigation sites (11.0) (linear contrast, P=0.0001). Diversity declined with age of the habitat (linear contrast, P=0.0001; new mitigation sites had a diversity of 12.6, old sites were 10.7, and natural sites were 8.0. The small difference between new ~nd old sites was significantly different (linear contrast, P=0.0453). The average diversity at each site follows (N1, station name, and tukey test): • 16.8 12.2 11.8 11.0 10.9 9.2 9.2 8.1 7.1 6.0 5.8 4.2 2.5 TSG SKG GIG 189G 189S CPG TPG TSS Pl1G GIS Pl2G GES 6 • Species evenness was different among stations (1-way ANOVA, P=0.0001, Figure 20, Table 9). The average E1 evenness index for seagrass beds was 0.80, which was not different from 0.77 for unvegetated stations (linear contrast, P=0.1173). Natural sites had a lower average evenness (0.70) than mitigation sites (0.82) (linear • contrast, P=0.0002). Evenness declined with age of the habitat (linear contrast, P=0.0003; new mitigation sites had an evenness of 0.83, old sites were 0.79, and natural sites were 0.70. The average evenness index at each site follows (E1, station name, and tukey test): • 0.97 0.91 0.88 0.85 0.83 0.83 0.80 0.77 0.76 0.74 0.68 0.62 0.57 TSS 189S TSG GIG GIS SKG 189G TPG CPG 6 Pl1G Pl2G GES • Evenness and diversity were correlated (Figure 21 ). As diversity increases evenness increases, i.e., dominance decrease-;s. There appears to be a phase shift, • or two separate relationships, for the non-vegetated sites versus the vegetated sites. The non-vegetated sites have higher evenness values than the vegetated sites. This indicates that there may be less dominance at n~n-vegetated sites . • 17 • Temporal Study • • Two paired stations, one natural (189) and one created (TS), were monitored for one year to determine if change in the newly created habitat were discernible. There was always more material in the natural sediments (189) than in the mitigation site (TS) (Table 10, Figures 22-23). The relative proportion of organic matter in the surface 0-3 • cm and bottom 3-10 cm sections of sediment did not change much over the year of monitoring (Figure 22). In general, the higher proportion of organic matter in the natural station (189) was due to higher amounts of material in both sections (Figure 22). At both sites, the organic material at the sand stations (-S) was composed entirely of detritus (Figure 23). There was much more detritus in the natural grass station (-G) than at the mitigation station (Figure 23). There was very little change at any station from April through October 1992. • • Seasonal fluctuations in macrofaunal abundance (Figure 24) and biomass (Figure 25) did occur. The interaction between stations and dates was significant for biomass (2-way ANOVA, P=0.0428) and abundance (P=0.0028), indicating that changes in abundance and biomass were different at the mitigation and the natural site. Abundance at the natural sites increased throughout the year, but at the mitigation site there was a large decline during the spring and then a rise for the remainder -of the year . • • • • • 18 • • • • • • Figure 3. Sediment composition. Per cent dry weight of sediment components in each station. Samples taken in April 1992 . • • •· • • • 19 • • • • • • • • • • • Laguna Madre Sediment Composition (% dry weight) Percent 100-,,-,.,.-,,.-...,.--.-~,..--r-~~~-n-:~-=-?TT-r~-=--r--n-~-,.,--n--77"-r-t7J'~~~--,.,.r--rr-»--"7r"T--i-7r-=-~~/7'1 90 80 70 60 50 40 30 20 10 0 189G 189S 6 CPG GES GIG GIS Pl1 Pl2 SKG TPG TSG TSS Station Clay ts