© Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at MYVATN-LAXA RAMSAR SITE - A CASE OF INTEGRATED MONITORING Ärni Einarsson Abstract Waterfowl populations at Lake Myvatn and River Laxä, Iceland, have been monitored since 1975. A hypothesis that the local populations of breeding waterfowl are determined by the food supply on the breeding grounds sti- mulated monitoring of the most abundant food species. Chironomid midges have been monitored since 1977. The record of two duck species, the Wigeon (Anas penelope) and Common Scoter (Melanitta nigra) and a critical food, chironomid midges, is used to illustrate how analyses of time lags can generate hypotheses about the causes of population changes in waterfowl. The records of waterfowl numbers, and also the catch of Brown Trout (Salmo truttd), suggest a close relationship with food conditions. A record of egg harvesting, going back to 1901, also indicates a close link with food conditions on the breeding grounds. By employing palaeolimnological me- thods it has been possible to gain a better historical perspective and assess the stability of the ecosystem. Einarsson, Ä.: Myvatn-Laxä RAMSAR Gebiet - ein Fallbeispiel für integriertes Monito- ring Monitoring von Wasservogelpopulationen wird am Myvatn-See und am Laxä-Fluß seit 1975 durchgeführt. Auf- grund der Hypothese, daß die lokale Brutpopulation der Wasservögel vom Nahrungsangebot an den Brutplätzen bestimmt wird, begann man mit Monitoringstudien an den abundanten Nahrungsarten. Seit 1977 findet Monito- ring an Chironomiden statt. Daten über zwei Entenarten, Pfeifente (Anas penelope) und Trauerente (Melanitta nigra) und über Chironomiden als Nahrung, sollen illustrieren wie die Analyse von Zeitverschiebungen zu Hypo- thesen über die Ursachen von Populationsschwankungen bei Wasservögeln führt. Angaben über Wasservogelbe- stände und Forellenfangergebnisse (Salmo trutta) lassen einen engen Zusammenhang zu den Nahrungsbedingun- gen vermuten. Daten über die Ernte von Eiern, die bis 1901 zurückreichen, zeigen ebenfalls enge Beziehungen zu den Nahrungsgrundlagen an den Brutplätzen. Durch palaeolimnologische Methoden wurde es möglich, eine bes- sere historische Einsicht zu gewinnen und die Stabilität des Ökosystems zu erfassen. Einarsson, Ä.: Ramsarska lokalita Myvatn-Laxä - pfiklad integrovaneho monitoroväni Populace vodniho ptactva na jezefe Myvatn a fece Laxä na Islandu jsou monitoroväny od roku 1975. Hypoteza, ze mi'stni populace hnizdfciho vodniho ptactva jsou zävisle na zdroji potravy na hnfzdistfch, byla podnetem k monitoroväni nejrozsffenejsi'ch druhü potravy. Populace pakomärü jsou monitoroväny od roku 1977. Analyzy populacnfch dat dvou druhü kachen hvfzdäka (Anas penelope) a turpana cemeho (Melanitta nigra) a jejich zakladnf potravy larev pakomärü jsou zäkladem hypotez o pffcinäch populacnfch zmen vodniho ptactva. Pocty vodnfch ptakü i ülovky pstruha (Salmo trutta) ukäzaly tesnou zävislost na potrave, stejne jako üdaje o sberu va- jec, ktere se datujf od roku 1901. Paleolimnologicke analyzy poskytly pfesnejsf historicke üdaje a tak umoznily zhodnocenf stability ekosysttmu. 211 © Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at INTRODUCTION monitoring. At the same time, all the now familiar questions about the purpose and Lake Myvatn is a large (37 km2), shallow, limitations of monitoring have emerged, - eutrophic lake in a region of active volca- if not in the beginning then in retrospect. nism in Iceland. The landscape is shaped In this paper I should like to use examples by volcanism and the present lake was from our monitoring programme and rela- created only about 2300 years ago by a ted research to illustrate how monitoring volcanic eruption which poured large vo- data can be used to gain insight into the lumes of lava over the district. The lake processes at work within the ecosystem. rests in a shallow depression in the lava- field. The lake and its outflowing river - the Laxä - have a worldwide reputation as AIMS AND PRINCIPLES a breeding and moulting place for water- fowl, especially ducks. Iceland is situated Monitoring has been defined in various on biogeographical crossroads, between ways (cf. GOLDSMITH, 1991 and North America and Eurasia, and the Arctic SPELLERBERG, 1991) but here I shall use a and the Boreal regions. This results in a definition which suits the Myvatn situati- unique species composition of waterfowl at on: "Monitoring is a long-sighted, syste- Myvatn. Larvae of Diptera constitute the matic and repeated documentation of popu- most important food for the ducks and the lations and their habitats." This definition local fish populations. In the lake chiro- is population oriented, rather than focusing nomids dominate, but in the river Simulium on environmental quality as displayed by is the dominant food species. The name chemical and physical parameters. "Myvatn" literally means "lake of midges" There are two main aims of monitoring. and reflects the large amounts of these flies One is to aid conservation, the other to on the banks (see JÖNASSON, ed., 1979). promote understanding of the ecosystem. The lake and the river were protected by These two aims interact: better understan- law in 1974, and in 1978 designated on the ding of the ecosystem helps conservation. RAMSAR list of wetlands of international Collection of data may, however, be done importance. Large changes had been ob- purely out of academic interest. The aims served in populations of both the vertebra- of monitoring can be defined further: tes and the invertebrates in the preceding years, and there were fears that recent in- a. To be an alarm bell, i.e. to give early dustrial development (mining of the lake notice that unwanted changes may be sediment for diatomite production, com- taking place. mencing in 1967) was harming the ecosy- b. To provide information on base line va- stem. A research station was established in riation in the ecosystem. 1974 and among its earliest aims was to c. To observe how perturbations travel monitor the animal populations. The cur- through the ecosystem in order to high- rent monitoring programme started with light hypotheses about causal links. bird censuses in 1975, but more animal d. To assess the efficiency of management populations were included later. The pro- measures. gramme developed hand in hand with our knowledge of the ecosystem which in- The aims may vary from place to place, creased considerably during the first years. depending on the characteristics of the eco- Academic interest in the lake is high and system, the type of threat, if any, and the interaction between the monitoring and political environment. The usefulness and academic research activities has played a limitations of monitoring for detecting en- crucial role in the development of the vironmental stress in natural populations 212 © Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at has been discussed by UNDERWOOD Why do these changes occur? At the outset (1989); see also CARPENTER (1988) for there were indications that recent changes discussions about temporal variability in in the duck populations were linked with freshwater ecosystems. food abundance in the lake. Older histori- cal data suggested that this might be the rule. Chironomid midges were known to be EXAMPLES OF MONITORING the most important food (BENGTSON 1971, GARDARSSON 1979). The Common Scoter, like most other diving ducks, takes the chi- When monitoring of animal populations ronomid larvae on the bottom of the lake or started in 1975 there were clear signs of a picks the flies from the surface when they poor situation in the waterfowl populati- emerge. The Wigeon has to rely on flies on ons. Waterfowl has a relatively high con- the water surface or larvae washed up on servation value, and seems suitable as an the shore with vegetation. Although the indicator of the general situation of the Wigeon is to a large extent a herbivore, area because there are many species with midges comprise an important source of different requirements. Also, waterfowl protein for the laying female and small can be counted rather easily. So when young. Midge larvae in Myvatn can be comparing cost and efficiency, waterfowl divided into two groups. There is a group was the obvious choice for starting a moni- of infaunal species, with larvae living in toring programme. As examples I should the bottom mud. These are the most abun- like to take the Common Scoter (Melanitta dant ones, represented by Tanytarsus and nigra) and the Wigeon (Anas penelope), a Chironomus species. The other group of diving duck and a dabbling duck (Fig. 1). species has epifaunal larvae, crawling upon The numbers in spring of these two species the bottom and moving about in vegetati- vary about 2-3 fold. Periods of increase on. The epifaunal larvae are preferred by may last 6-8 years, decreases seem to take the diving ducks, probably because of a shorter time. The variation is large greater accessibility. enough to conceal a long term trend, if it exists. 2000H 600 MELANITTA NIGRA 500 1600- CO LLI < 400- AA 1200- u. w 0 300- tr m 600- s 1 200 400- 100- 0 1975 1980 1985 1990 1975 1980 1985 1990 YEAR YEAR Fig. 1.: Numbers of male Wigeon (Anas penelope) and Common Scoter (Melanitta nigra) in spring in the Lake Myvatn area. 213 © Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at Monitoring of midges started in 1977. As a dependent variable, we use the change Specially designed window traps have be- that had occurred in duck numbers since en employed to catch flying insects the previous year. As independent varia- (JONSSON et al. 1978). The traps are emp- bles, we use the catch of chironomid mid- tied at 7-10 day intervals in the summerti- ges in the year (t) and the production of me. About 50 different species of chiro- young in the previous year (t-1), two years nomid midges have been recorded, about earlier (t-2) and three years earlier (t-3). 10 of which play a significant role in terms Analyzing longer time lags is hardly worth of abundance. The total catch of chirono- while in this case because too many data mid midges (Fig. 2) in 1977-89 shows points will be lost. fluctuations which suggest a link between The duck populations studied did not show the food supply and the duck populations significant correlations with the numbers (GARDARSSON & EINARSSON, in press). of chironomid midges in the same year. We wanted to consider the question whe- This means that in spring the ducks do not ther the duck numbers respond directly to react directly to the food supply. Both the the food situation or whether the response Common Scoter and Wigeon numbers cor- is through the production of young. In related with the number of young produced other words: do duck numbers increase the year before. In the case of the Wigeon because adults become attracted to richer this is understandable because yearlings food supplies or because they produce mo- return to the breeding grounds. The re- re young? By analysing time lags between sponse of the Common Scoter is more sur- changes in the food situation and changes prising because their young do not return in the duck populations we may learn to the breeding grounds until later. The which process is the more important. If the one- year lag suggests that the adult birds response is slow (long time lags) it be- "remember" the situation from last year. comes more likely that the population is Good food conditions the previous summer reacting by increased production. Multiple may encourage the birds to return to regression analysis makes it possible to Myvatn rather than going somewhere else. analyze the importance of several indepen- Perhaps they simply survive better. dent variables simultaneously. 25<h 1CH 83 8^ S * < 4- O • f 2 V (H-r- 1978 1980 1982 1984 1986 1988 1978 1980 1982 1984 1986 1988 YEAR YEAR Fig. 2: Catches of adult chironomid midges in flytraps at Lake Myvatn. A: all chironomids; B: chironomid midges whose larvae are epifaunal. (Data from GARDARSSON & EINARSSON, in press) 214 © Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at 3.0n Anas penelope Melanitta nigra r = 0.80 " r = 0.89 •••• 4.0 2.5- o 3.5 2.0 i o u. O 3.0- <r 1.5- m I D Z 2.5- 1.0 2.0 0.5 3.5 4.0 4.5 5.0 5.5 3.5 4.0 4.5 5.0 5.5 NUMBER OF RJES (log) NUMBER OF FUES (log) Fig. 3. Production of Wigeon (Anas penelope) and Common Scoter (Melanitta nigra) young in relation to the total catch of chironomid midges in fly traps. The data point in parenthe- ses is excluded from the analysis. It is from 1983, when counting methods differed and re- sulted in an overestimation of production. (Data from GARDARSSON & EiNARSSON, in press) The young of the Common Scoter do re- the previous year indicates a significant turn later, and a significant correlation oc- relationship (Fig. 4) (GISLASON, in press). curs between the change in spring numbers and the production of young three years earlier (GARDARSSON & EINARSSON, in 5000n press). We are aware of the pitfalls involved in correlating time series and also in relying too heavily on the results of multiple re- gression analyses. Our intention is to sort out the relationships which seem to be the most significant in order to get a better focus for future research. The link with food can be further clarified by analyzing the relationship between the food situation and the production of young. The relationship is statistically significant 50 100 150 200 250 300 for both species (Fig. 3). THOUSANDS OF FLIES The Brown Trout (Salmo trutta) in the Ri- Fig. 4.: Catch by rod fishing of Brown ver Laxä is another example of how moni- Trout (Salmo trutta) in the River Laxä toring of different trophic levels can be in relation to the previous year's catch of useful. Catch statistics (from angling) are blackflies (Simulium) in flytraps. available from 1976 onwards. The main (Redrawn from GISLASON, in press). food of the Brown Trout, the blackfly (Simulium vittatum), has been monitored The relationship is curvilinear: the total with fly traps since 1977. A regression of catch of Brown Trout seldom exceeds 4500 the catch of trout on the food situation in fishes. This is probably because of catch 215 © Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at restrictions, or alternatively, because at diatoms, chrysophyceans as well as plant high food levels other environmental fac- pigments (e.g. HA WORTH & LUND, 1984; tors than food may become limiting. GRAY, 1988). This kind of information is In this example we can go further down the useful in revealing the conditions in a food chain. The blackfly larvae are filter wetland before, e.g. the industrial revoluti- feeders and planktonic algae (mostly Ana- on or forest clearance, and also gives an baena flos-aquae) and detritus drifting indication of the long term stability of the from the lake comprise their main food. ecosystem. Some of the microfossils in the The drift of algae varies considerably bet- sediment are derived from species which ween years. A multiple regression analysis are important in the food web, such as the reveals a significant relationship between chironomid midge Psectrocladius barbi- blackfly numbers and both drift of Ana- manus and the cladoceran Eurycercus la- baena and water temperature (GiSLASON, mellatus (Fig. 6). In Lake Myvatn, both in press). species are associated with filamentous, mat-forming green algae (Cladophora ae- gagropild) on the bottom of the lake. Pa- laeolimnological studies indicate that in HISTORICAL PERSPECTIVE Lake Myvatn these species have been more common in the last 270 years than earlier Very few biological monitoring program- in the history of the lake, which spans mes have been running for more than about 2300 years. This has been linked twenty years. This number of data points is with a corresponding increase in Clado- minimal if we want to detect long term- phora (EINARSSON & HAFLIDASON, 1988; trends, given the variation in the data. A EiNARSSON et al., 1993). search for historical data may be fruitful. Most often such information is based on harvest figures: fishing or hunting, and can be difficult to interpret because of varying DISCUSSION and often unknown intensity of harvesting. At Myvatn we have fishing data from 1900 Monitoring is, strictly speaking, an endless to the present day, also figures about har- process and because of financial con- vesting of duck eggs from 1901 to 1957 straints we have to be selective about what (GUDMUNDSSON, 1979). The similarity of to monitor and about the methods em- the harvest figures of Common Scoter ployed. We have to make efforts to combi- {Melanitta nigra) and Long-tailed Duck ne low cost and high efficiency, and have (Clangula hyemalis) eggs is interesting to be certain that our methods will be valid (Fig. 5) (see GARDARSSON et al. 1988). for a long time. At the same time, we The similarity cannot be explained by should make efforts to identify the most changes in harvesting practice. The two important factors in the ecosystem, and I duck populations have different wintering have demonstrated how a monitoring pro- grounds, and these data indicate that chan- gramme may be useful in this process, al- ges in duck numbers at Myvatn are pri- though by no means sufficient. I would like marily governed by local processes rather to stress the importance of having well than processes operating on the wintering formulated hypotheses, preferably more grounds (GARDARSSON, 1979). than one, about the crucial factors deter- The historical perspective can be improved mining the populations in question, be it considerably by employing palaeolimno- animals or plants. Equally important is to logical methods. A number of techniques formulate hypotheses about the impact of are available, which allow tracing the hi- man-induced changes. story of, e.g., chironomids, cladocerans, 216 © Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at 3000-] Melanrtta nigra Clangula hyemalis 2500- g 2000- u. o a: 1500-] UJ 03 = 1000- 500- 1900 1910 1920 1930 1940 1950 1960 Fig. 5. Harvest of eggs of Long-tailed Duck (Clangula hyemalis) and Common Scoter (Melanitta nigra) at Myvatn 1901-1957. Three years running averages. (Data from GUDMUNDSSON, 1979). Psectrocladius barbimanus DEPTH (m) IN SEDIMENT Fig. 6.: A palaeolimnological record of 2300 years of the chironomid Psectrocladius and the cladoceran Eurycercus from the Lake Myvatn sediment. The location of two dated volca- nic ash layers (A.D. 1477 and A.D. 150) is also shown. The densities are thousands of fragments per g organic sediment. Redrawn from EINARSSON & HAFLIDASON (1988). 217 © Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at The problem most of us are faced with is by interested scientists, carrying out well the lack of baseline data. The need for designed research projects. monitoring a piece of wetland does not The food web is a fundamental feature of become generally recognized until after the the ecosystem. When unexpected changes threat is there and after symptoms have occur in an animal population, changes in emerged. We normally do not know the so- the food situation should be suspected. called natural variation in the ecosystem or Monitoring different trophic levels at Lake the variation that existed before the most Myvatn and River Laxä has been especial- recent threat was realized. The frequent ly rewarding in this respect. Statistically lack of control is also frustrating: quite significant relationships within the food often we are dealing with some precious web, like the ones I have presented here, piece of wetland, and there is no other La- are stimulating, because they suggest that ke Myvatn to compare with. And even we might be on the right track. But the lack though there was a similar lake, because of of significant relationship between two cost considerations we have to choose time series is equally important, especially whether to spread the effort and run a in the light of hypothesis testing. Food web parallel monitoring programme in the con- studies are gaining more attention because trol lake or whether we should concentrate of recent crystallization of theories about on our own lake that we know so much the controlling factors of lake communities about already. There is no golden rule. ("top down" and "bottom up") (see Each case is in some way special and what KERFOOT & SIH, 1987; MCQUEEN et al., we do depends on the local situation and 1989). I would not like to say that we the hypotheses posed. should concentrate all our effort on food Each wetland has its characteristics and webs. By doing so we may miss other im- ecological principles can only be applied in portant parameters. But omitting the study a rather general way. Therefore we must of the food web seems to be unwise. formulate hypotheses about the dynamics of the specific wetland we are interested in. ACKNOWLEDGEMENT Hypotheses, as a tool of science, have an important property: they have to be falsi- Prof. Arnthor Gardarsson read the ma- fiable, and furthermore, every effort should nuscript and suggested a number of impro- be made to prove them wrong. What then vements which are gratefully acknowled- is the value of letting hypotheses govern ged. our choice about what to monitor if we run the risk of spending time and money for REFERENCES years in watching a parameter which may prove unimportant? The answer lies in the long-term objective. Only by understan- BENGTSON, S. A. (1971): Food and feeding ding the functioning of the ecosystem can of diving ducks breeding at Lake we hope to be able to deal properly with Myvatn, Iceland. Ornis fennica 48: 77- future threats. Only by rejecting hypothe- 92. ses can we drop some parameters from our CARPENTER, S.R. (ed.) (1988): Complex monitoring programme and make it chea- Interactions in Lake Communities. per. All hypotheses about function have to Springer Verlag. New York. 283 pp. take into account the long term variation, EINARSSON, Ä. (1982): The palaeolimno- or stability, of the parameters involved, and logy of Lake Myvatn, northern Iceland: this is one reason why monitoring becomes plant and animal microfossils in the se- an important practice. This process may diment. Freshwater Biology 12: 63-82. take a while and needs constant attendance 218 © Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at EiNARSSON, Ä. & H. HAFLIDASON (1988): JONSSON, E., A. GARDARSSON & G.M. Predictive palaeolimnology: Effects of GISLASON (1986): A new window trap sediment dredging in Lake Myvatn, used in assessment of the flight periods Iceland. Verh. der Internat. Vereinigung of Chironomidae and Simuliidae für theoretische und angewandte Limno- (Diptera). Freshwater Biology 16: 711- logie 23: 860-869. 719. EiNARSSON, Ä., H. ÖSKARSSON & H. KERFOOT, W.C. & A. Sm (eds.) (1987): HAFLIDASON (1993): Stratigraphy of Predation. Direct and Indirect Impacts fossil pigments and Cladophora and its on Aquatic Communities. University relationship with deposition of tephra in Press of New England. Hannover and Lake Myvatn, Iceland. Journal of Pa- London: 386 pp. leolimnology 8: 15-26. MCQUEEN, D.J., M.R.S. JOHANNES, J.R. GARDARSSON, A. (1979): Waterfowl popu- POST, T.J. STEWART & D.R.S. LEAN lations of Lake Myvatn and recent (1989): Bottom-up and top-down im- changes in numbers and food habits. pacts on freshwater pelagic community Oikos 32: 250-270. structure. Ecological Monographs 59: GARDARSSON, A., G.M. GISLASON & Ä. 289-309. EiNARSSON (1988): Long term changes SPELLERBERG, J.F. (1991): Monitoring in the Lake Myvatn ecosystem. Aqua Ecological Change. Cambridge Univ. fennical8: 125-135. Press. Cambridge UK: 334 pp. GARDARSSON, A. & A. EINARSSON (1994): UNDERWOOD, A.J. (1989): The analysis of Responses of breeding duck populations stress in natural populations. Biological to changes in food supply. Hydrobiolo- Journal of the Linnean Society 37: 51- gia, in press. 78. GiSLASON, G.M. (1994): River manage- ment in cold regions. A case study of the River Laxä, North Iceland. In: P. CALOW & G.E. PETTS (eds.), The Rivers Handbook Vol. 2.: 464-483. Blackwell Scientific Publications, Oxford. GRAY, J. (ed.) (1988): Aspects of freshwa- ter paleoecology and biogeography. Palaeogeography, Palaeoclimatology, Palaeoecology 62: 623 pp. GUDMUNDSSON, F. (1979): The past status and exploitation of the Myvatn water- fowl populations. Oikos 32: 232-249. HAWORTH, E.Y. & J.W.G. LUND (eds.) (1984): Lake Sediments and Environ- Address of the author: mental History. Leicester University Press: 411 pp. Ami Einarsson HELLA WELL, J.M. & F.B. GOLDSMITH (eds.) (1991): Monitoring for Conserva- Myvatn Research Station tion and Ecology. Chapman and Hall: c/o Institute of Biology 273 pp. University of Iceland JONASSON, P.M. (ed.) (1979): Ecology of Grensasvegur 12 eutrophic, subarctic Lake Myvatn and 108 Reykjavik the River Laxä. Oikos 32: 1-308. ICELAND 219