Limnologisches Institut, Universität Konstanz. Konstanz, Germany Susceptibility ofLakeside Phragmites Reeds to Environmental Stresses: Examples from Lake Constance-Untersee (SW-Germany) WOLFGANG OSTENDORP With 3 Figures and 2 Tables Key words: Lake shore vegetation, Phragmites australis, die-back of reeds, mechanical stress, flooding stress, model Abstract The lakeside border of Phragmites reeds is under the control of models which describe the susceptibility of Phragmites to many environmental factors, among which the flooding height and mechanical and flooding stress are discussed. Most ofthe de­ the degree of mechanical load by waves and washes are the most tails have been published elsewhere (e.g. OSTENDORP 1990, prominent. Examples from Lake Constance-Untersee show that the 1991 b, 1992, 1995 a). reed front retreats under enhanced stresses. Models are presented that describe the susceptibility of lakeside reeds to mechanical dam­ age and to flooding events. A monitoring programme which ran from 1984/85 to 1992/93 showed a slow lakeward progression of the Investigation area reeds in a decade of abnormally low water conditions. The investigations were performed at Lake Constance-Un­ tersee, the shallow eutrophic lower lake basin of Lake Con­ Introduction stance, with extended lake shores (10.2 km2 littoral area be­ tween 394.0 and 396.0 m a.s.l., mean width 143 m, mean The Common Reed, Phragmites australis (CAV.) TRIN. ex slope angle 0.80°). 99.7% of the fringing reeds consists of STEUD. (Poaceae) is a world-wide distributed grass with a (nearly) monospecific Phragmites australis stands, covering wide ecological amplitude (RODEWALD-RuDESCU 1974; CON­ a total of 295 hectares (German territory only). The sublit­ ERT 1983). At the shallow-sloped shores of many European toral zone is overgrown with submerged macrophytes, main­ lakes it forms monospecific belts which often cover many ly Chara spp., Potamogeton spp. and filamentous algae (e.g. hectares, or even square kilometres. The landside border of Cladophora sp.), which are often uprooted by waves during pure Phragmites reeds is controlled by the competitive power storms in autumn, and washed onto the shore (OSTENDORP of trees and bushes, and by human impacts (wet meadow use, 1991 a, 1992). construction of lake walls, land fills, etc.). The factors that The yearly course of the lake level of Lake Constance determine the lakeside reed front are much less dear: pre­ mainly depends on the discharge of the Alpenrhein river, so sumably a specific combination of mean and/or maximum that in lune and luly the water level reaches its maximum, flooding height, or flood frequency, mechanical stress by and it drops down gradually to the yearly minimum in Febru­ waves and drifting matter, grazing by coypus, muskrats and ary and March (mean water level = 395.09, mean low water waterfowl, and sediment quality (e.g. redox conditions and = 394.35, mean high water = 396.23 m a.s.l., period oxygen consumption capacity) plays an important role 1937-1987, water gauge station Berlingen, LUFT et al. (CooPS et al. 1994, 1996; WEISSNER 1987,1991). 1990). The annual course of the water level is modified by The importance of these factors becomes obvious by episodic high flooding events which occur after heavy rain­ analysing extreme events which have led to a destabilization fall in the prealpine mountains. Flood disasters with water of the lakeside reed front and to an extensive reed dedine, as levels higher than 396.65 m a.s.l. have frequently happened it was the case at Lake Constance. In this paper a concise during this century, but they have been less frequent in the summary of the factors of reed die-back is given, and two last decade (for the last time in 1987; DIENST 1994). Normal- 0075-95ll/99129/01-21 $ 12.00/0 21 / Iy, they are restricted to the period from June to August cov­ shore seetions with a high spatial and temporal resolution ering the second half of the growth period of Phragmites (PIER et al. 1993; OSTENDORP et al. 1996). 10 sets of data are australis. Westerly winds prevail changing from WNW in currently available comprising up to 51 test seetions of 25 m spring and summer to WSW in winter. Strong winds and length each. The data demonstrate that many lakeside reeds storms are rare and occur mainly from October to April, in have spread anew since the middle of the 1980s. The high seasons when the water level is low. progression rates between 1984/85 and 1987/88 with 4.3 to 15.5 em a- 1 were followed by a regression of --4.8 cm a- I on average in 1988/89. Since that time a eontinous advance has Dynamics of the lakeside reeds at Lake been observed with mean rates between 11.4 und 49.2 cm Constance-Untersee a- 1• For wind and wave exposed stands individual progres­ sion rates were lower than the average in all years of all In the late 1960s a serious decline of lakeside reeds was stands. effected by professional fishermen. This was the starting­ The data of this monitoring programme and of the dra­ point of aseries of investigations into the causes and circum­ matie decline around 1965 1ed to the hypothesis that me­ stances of the die-back (SCHRÖDER 1979, 1987; OSTENDORP chanieal stresses to the 1akeside reed front and the maximum 1990, 1991 a, b, 1992). After decades of more or less contin­ flooding height during summer are among the most impor­ uous lakeward progression the decline began at severalloca­ tant factors determining the dynamics ofthe fringing reeds in tions between 1954 and 1962. By far the strongest decline Lake Constance-Untersee. happened between 1962 and 1967, so that by 1967 18% of the fringing Phragmites reeds (69 hectares, 1954 = 100%) were lost. In the following years another 5% died back Significance of extreme flooding events (Table 1). The outermost reed stands (394.0-394.5 m a.s.l.) were the most affected (-75%), whereas the landside fringe The rhizomes of1akeside reed stands are buried in the water­ (395.5-396.0 m a.s.l) lost only 19% of its area by 1954. 10gged sediment. Henee, in spring and early summer the Due to the lakeside losses the mean depth penetration young shoots must grow through a water column of varying line of Phragmites shifted from 394.49 ± 0.13 m a.s.l. to height. By its architecture, Phragmites australis is a terrestri­ 394.95 ± 0.22 m a.s.1. A detailed evaluation of the temporal al plant which lacks any adaptations to submersed carbon as­ pattern of the decline showed that it was the summer of 1965 similation. Consequently, the photosynthesis will be arrested whieh was characterized by extreme meteorologieal and hy­ if the leaves are waterlogged (RODEWALD-RuDESCU 1974: drologieal events. The water level in June was the fifth high­ 81). The culms react on prolonged submergenee with an ac­ est since the beginning of water mark reeords in 18 l7, and eelerated growth, but finally die as soon as the earbohydrate events with squalls and strong winds of 6° Beaufort and reserves in the rhizomes are exhausted. Additionally, the lim­ more were by 36% more frequent than on average in the pe­ itation of the oxygen supply to the rhizomes by the sub­ riod between 1959 and 1984. merged green parts of the culms may play an important role In the winter of 1984/85 a monitoring programme was since the anoxie glucose metabolism yields by far a lower started to reeord the reed front dynamies at representative energy gain than the oxie respiration (BRÄNDLE & CRAW- Table 1. Decline offringing Phragmites reeds at Lake Constance-Untersee 1954-1983: reed bed area lost in the period 1954/c. 1983. Total reed bed area was 265 hectares (= 87% of the littoral area) for the landside reeds in 1954, and 114 hectares (= 18% of the littoral area) for the lakeside reeds. Period 1954-1964 1965-1967 1968-1972 1972-1978 1978-1983 Landside reeds: Decline 36 ha (41 %) 8 ha (9%) oha (0%) (?) Main causes land fills. boating places, land fills, boating places, ? camping sites etc. camping sites etc. Lakeside reeds: Decline 10 ha (l1%) 23 ha (26%) 4ha(5%) 4ha(5%) 3 ha (3%) Main causes ? flood disaster, mechanical damage by winter harvesting storms, hailstorm macrophyte wash, and burning in summer 1965 bank erosion ofreeds 22 Limno10gica 29 (1999) 1 FORD 1987; BRÄNDLE 1991). Therefore, it is assumed that According to this model the following situation was recon­ growing Phragmites shoots must not be completely flooded structed: In the first two months ofthe growth period (from the for longer than 3 days, or must not be flooded higher than to beginning of May to the beginning of luly) in 1965 about two a critical culm length Leri" so that at least two expanded thirds of lakeside culms were submerged over several weeks leaves are left unflooded. (W > Lcri' + Wo), This must have led to a die off of all these The degree of threat to lakeside Phragmites reeds by culms. Additionally, the flood peak at the end of lune was ac­ floods can be estimated if the following parameters are companied by strong winds and hail storms so that many ofthe known: (1) the criticallength (Lcri,) as a function of time t, (I1) taller culms which had successfully competed with the rising the water mark (W) as a function of time, and (lll) the eleva­ water level were destroyed. In 1987 the maximum water level tion of the ground of the Phragmites stand relative to the was only a few centimetres below ofthat of 1965, but the flood zero-mark of the water gauge (Wo): began to swell three weeks later.