Aarhus University Seal / Aarhus Universitets segl

Karsten Dahl

Test af metoder til marine vegetationsundersøgelser: Test of Methods for Investigations of Marine Vegetation

Publikation: Bog/antologi/afhandling/rapportRapportRådgivning

  • Afdeling for Marin Økologi
English summary (half way down) Baggrund Behov for nye metoder til vegetationsundersøgelser Den marine vegetation er blevet undersøgt mere eller mindre in-ten-sivt i mere end 100 år. Forskellige undersøgelsesmetoder har været anvendt. Generelt er de fleste undersøgelser før 1940 udelukkende kvalitative (artslister), mens de senere undersøgelser i stigende grad er kvantitative. Selvom metoderne er blevet forbedret, indgår der ofte subjektive bedømmelser, hvilket giver anledning til stor variation i de undersøgte parametre, når forskellige dykkere foretager under-søgelserne (Middelboe et al., 1997). Mere reproducerbare kvantitative metoder er vigtige for at kunne sammenligne undersøgelser foretaget forskellige steder eller på forskellige tidspunkter. Med nye forbedre-de metoder vil man ikke blive i stand til bedre at sammenligne med tidligere undersøgelser, men grundlaget for sammenligning mellem undersøgelser fra forskellige steder bliver bedre. Man vil også kunne forbedre grundlaget for fremover at vurdere de tidslige ændringer i artssammensætning og udbredelse af den marine vegetation for der-igennem at vurdere effekter af menneskeskabte forstyrrelser på ve-getationen. De eksisterende metoder To ofte anvendte metoder til undersøgelse af den marine vegetation er punktundersøgelser og transektundersøgelser (Krause-Jensen et al., 1995 og 1998). Metoderne anvendes i dag i det nationale overvåg-ningsprogram, NOVA, hvor vegetationen undersøges på hård og blød bund i de kystnære områder samt på en række stenrev. I et punkt eller i dybdeintervaller langs et transekt bestemmer en dykker artssammensætningen og arternes dækningsgrad. Metoderne er dog ikke optimale, idet interkalibreringer har vist, at der kan være store variationer i vegetationens sammensætning og meget store forskelle mellem bestemmelserne af arternes dæknings-grad, når forskellige dykkere har undersøgt det samme punkt eller transekt (Middelboe et al., 1997; Dahl et al., 2000). Interkalibreringerne har derfor rejst spørgs-målet om, hvorvidt der findes andre, mere velegnede metoder til at undersøge den marine vegetation end de punkt- og transekt-under-søgelser, der i dag anvendes i det nationale overvågningsprogram. Subjektivitet på flere niveauer Det største problem i de eksisterende metoder er, at de er subjektive på flere niveauer. Valget indgår dels i defineringen af undersøgelses-arealet, dels i vurderingen af arternes dækningsgrad i synsfeltet, og dels i vurderingen af arternes gennemsnitlige dækningsgrad i under-søgelsesarealet. Hverken ved punkt-undersøgelser eller transektun-dersøgelser har det undersøgte areal været veldefineret. Vegetatio-nen undersøges i et område omkring punktet og langs transektet, men den egentlige radius eller bredde af området defineres af den enkelte dykker. Dykkerens vurdering af dækningsgraden vanskelig-gøres desuden af, at dækningsgraden ofte skal estimeres inden for et stort areal, som dykkeren ikke umiddelbart kan overskue. Ved tran-sektundersøgelser skal dækningsgraden vurderes inden for dybde-intervaller, som kan have en dybdeforskel på op til 2 meter. Dæk-ningsgraden kan ændre sig betydeligt inden for et dybdeinterval, og det kan derfor være vanskeligt at estimere en gennemsnitlig dæk-ningsgrad for intervallet. De metodiske problemer er typisk større for den diverse vegetation på hård bund end for vegetationen på blød bund, der i de danske kystområder ofte er en monokultur af ålegræs (Middelboe et al., 1997), eller i nogle lavvandede bugter og fjorde af blandede bevoksninger af bl.a. Ruppia -, Potamogeton -, Zannichellia - og Chara. English summary Summary Three series of investigations were performed in order to assess methods of optimising monitoring of benthic marine macrophytes in Danish waters. The investigations focused on qualitative and semi-quantitative analyses of different macroalgal and eelgrass communi-ties. The tests were performed on three different types of communi-ties in three different habitats: 1) complex macroalgal communities on a stone reef in open waters, 2) less complex, nearshore macroalgal communities on hard substrate (stones) in a fjord and 3) eelgrass beds on soft and sandy substrates in a fjord. Macroalgal communities on stone reefs The reef investigation aimed at developing a sampling strategy to de-scribe temporal and spatial variation in macroalgal communities. The tests measured precision of the methods and reproducibility among observers. Sampling was performed to identify the occurring species and estimation of percentage cover of each species. Individual stones were used as sampling units and one sampling included 8-10 stone replicates. In order to reduce sampling variation, a stone size of 30-40 cm was used throughout the tests. The samples were analysed with multivariate statistics. This method allowed statistically significant vertical separation of macroalgal communities with a resolution of 1 m depth, and identified significant temporal changes over a period of 5 weeks. On greater depths, it also identified significant differences between neighbouring locations. The tests showed significant differ-ences between divers, but intercalibration on location reduced this difference to an acceptable level. Estimation of the species number on the stone reefs was highly variable and therefore with the present ef-fort the method does not give a reliable measure of occurrence on stone reefs. The great variation in species number was ascribed to a community structure with a high proportion of rare species on stones of different sizes. Nearshore macroalgal communities Five different methods of estimating species number and abundance were tested and compared. The methods differed with respect to sampling area (0.3-25 m²) and/or quantified species abundance, either by visual estimates of area cover or by frequency counts of pres-ence/absence. Four divers tested all methods at each of 4 marked lo-cations (25 m²). The methods were compared and evaluated with re-spect to reproducibility, precision and costs. The variations between methods and divers were analysed using a mixed ANOVA-model. Generally, frequency counts were less precise, less reproducible and required more resources than visual cover estimates. Differences among methods in precision and reproducibility of cover estimates were, however, only significant in few cases, and the differences be-came non-significant when cover was expressed in relative units as opposed to absolute units. Coefficients of variation were relatively large, implicating that many subsamples were needed to obtain satis-factory results. The number of species found by the five methods also differed significantly, with species number depending directly on sampling area. Overall, we concluded that the best results were ob-tained using visual cover estimates in a large area (25 m2). The spatial variation in macroalgal communities was subsequently tested at different scales in a fjord in order to define an optimal sam-pling strategy with a given resource available. In this investigation, we applied the method based on visual cover estimates in a large area (25 m²). Species composition differed markedly among the inner and outer fjord area. The outer fjord showed larger variability within sites as compared to among sites. Cost-benefit analysis consequently recommended a sampling strategy involving few sites with many subsamples per site in the outer fjord. By contrast, the inner fjord showed similar variability within and among sites, and the sampling strategy was optimised by using many sites with few subsamples per site. Seagrass beds This part of the investigation focused on parameters describing the distribution of seagrasses. All tests were done on eelgrass beds, but the methods also apply to beds of other seagrasses. We tested and compared four different methods of estimating percentage eelgrass cover. Method 1 included diver-reported estimates of average cover at depth intervals of 1 m. With method 2, the diver continuously re-ported any change in eelgrass cover, and for each observation, depth and position were recorded. Method 3 was similar to method 2 ex-cept that the diver reported observations at regular intervals defined by the tender. Finally, method 4 involved continuous video record-ings along the transects and subsequent evaluation of percentage cover at intervals of 5 seconds on the video, which approximately corresponded to an observation for every 2-5 m along the transects. Two divers tested all methods at each of the 5 transects situated per-pendicularly to the coastline. The results were analysed with respect to reproducibility, precision and costs using a mixed ANOVA-model. The results showed that methods 2 and 3 were very similar, but method 3 gave the most reproducible and precise results at the low-est cost. Observations based on video recordings were also precise but more time consuming. The video, on the other hand, had the ad-vantage of serving as documentation. An analysis of spatial variation in eelgrass cover showed greater variability among depth intervals than among transects, which im-plies that the former is decisive for the amount of transects that can be investigated for a given resource. If the resource is fixed to the present level of the Danish monitoring programme (NOVA 2003) the optimal sampling strategy in a typical Danish fjord would be 4-15 transects per area depending on the local variation between depth intervals. The investigation also showed that estimated depth limits of eelgrass should be based on 7-29 observations depending on the local variation in order to obtain a confidence interval of 0.5 m. Based on previously collected data, an analysis of temporal variation con-cluded that with 4-15 transects it was possible to identify a 2-10% change in percentage eelgrass cover per year during a 5-year period. Finally, we compared estimates of eelgrass area cover based on aerial photographs and a newly developed GIS-interpolation method. With the new GIS-interpolation method we could reduce the required number of observations considerably as compared to earlier interpo-lation methods. Aerial photographs and GIS-interpolations gave similar estimates of average area cover. The interpolation method might become useful in areas of limited size where the interpretation of aerial photographs is jeopardised (e.g. low contrast between vegetated and non-vegetated areas, mixed vegetation, banks of blueshells, reduced secchi depths).
OriginalsprogDansk
ForlagDanmarks Miljøundersøgelser, Aarhus Universitet
Antal sider120
StatusUdgivet - 2000
SerietitelFaglig rapport fra DMU
Vol/bind323

Se relationer på Aarhus Universitet Citationsformater

ID: 12471970