Examensarbete i biologi avseende kandidatexamen, 15 hp
VT 2019
THE EFFECTS OF
FORESTRY ON STREAM ECOLOGICAL INTEGRITY
Edith Bremer
The effects of forestry on stream ecological integrity Edith Bremer
Abstract
This study investigates the effects of forestry on leaf litter decomposition in small forest streams. Riparian forest, that is the land closest to the stream, maintain shading, water temperature and energy supply through litter fall. If the riparian zone is deforested, many riparian functions important for the integrity of the stream ecology, hydrology and
biogeochemistry can be lost or modified. Leaf litter decomposition can be used as an
integrated measure of the physical and biological changes following forestry perturbations.
This study was conducted in 11 northern and 12 southern Swedish streams to address; 1) How is leaf litter decomposition in small streams affected by forestry by measuring leaf litter decomposition in streams with different buffer widths, and; 2) How other environmental variables, such as stream bottom substrate, canopy openness, water temperature and stream velocity affected leaf litter decomposition. Buffer width had no effect on decomposition.
Temperature and proportion organic bottom substrate had respectively positive and negative trends with decomposition in the southern Swedish sites which suggests the importance of forestry targeting these riparian functions especially when managing small streams. At the northern sites, velocity showed a positive, and temperature a negative trend with leaf litter decomposition but none of these were significant. It is possible that the extraordinarily warm and dry weather before and during the study was conducted affected aquatic organisms to the degree that decomposition was inhibited, and most trends became too small to detect or that buffer width is less important in a warmer climate.
Key words: leaf litter, decomposition, forestry management, riparian buffers, headwaters.
Table of contents
1 Introduction and background ... 1
1.1Aims of the study and scientific questions ... 2
2 Materials and methods ... 2
2.1 Study sites ... 2
2.2 Field sampling ... 3
2.3 Litter decomposition ... 3
2.4 Statistical analyses ... 3
3 Results ... 4
4 Discussion ... 7
References ... 9
1
1 Introduction and background
In 1975, Noel Hynes wrote that in every respect, the valley rules the stream, meaning that every stream is part of a valley and affected by it. Especially the riparian zone, i.e. the strip of land surrounding watercourses, has a large effect on the in-stream processes, including ecology, hydrology and biogeochemistry (Naiman and Decamps 1997). The presence of vegetation around a stream reduces surface runoff to the stream through evapotranspiration.
The roots prevent excessive amounts of nutrients from reaching the stream and increase stability in the bank preventing erosion. The tree canopies regulate temperature through shading and stems and shrubs slow down and retain water, sediment and organic material during floods and overland flow (Sweeney et al. 2014, Zinko 2005). The smaller the stream, the more important the riparian zone becomes. Among other things, a larger portion of the water in the small streams comes from the groundwater flowing through the riparian zone.
Further, the canopy above small streams is often enclosed completely, thereby inhibiting primary production which in turn gives the decomposition of leaf litter a major role for energy transfer to consumers in local and downstream food webs (Allan 1995). Complete removal of forest cover around streams disrupts the riparian functions and thereby causes degradation of the water and habitat quality and the biota of the stream (Sweeney et al.
2004). Managers can protect these functions by refraining from cutting all the way down to the water and instead leaving a buffer zone of trees along the stream, however more
knowledge is needed to determine how these buffers should be designed to maximize ecosystem functions (Richardson et al. 2012).
A number of different indicators can be used to asses ecological integrity in streams and their reaction to perturbation. A few examples are taxonomic richness in fish and invertebrates or presence of certain sensitive species, ecosystem respiration, biogeochemical processes, primary productivity and decomposition (Clappcott et al. 2011). Leaf litter decomposition is an especially useful tool when assessing ecological integrity and the effects of environmental changes in small streams (e.g. Lidman et al. 2017, Erdozain et al. 2018). This is due to its fundamental role as an ecosystem process in these streams. Detritivorous aquatic insects time their growth to seasonal peaks in litterfall and then together with microbes and fungi facilitate the transformation, cycling, and downstream transport of carbon and nutrients in river networks (Lidman et al. 2017). The process of litter decomposition is an integrated measure of stream physical and biological components which also change in their influence over time (Lidman 2017). The decomposition of leaf litter is affected by a wide range of variables such as temperature, stream velocity, water chemistry, invertebrate fauna and light conditions, all of which are affected by human land use, forestry included (Allan 1995, Lidman et al. 2017, Sweeney 2004). The practice of intensive forest harvesting in Sweden may therefore affect litter decomposition thought number of in-stream changes, either biological or physiochemical.
It is especially interesting to study what variables affect leaf litter decomposition in a Swedish forestry context. Leaving buffers around streams when harvesting is used as a tool to protect the stream. To assess the effectiveness of such measures we can potentially use leaf litter decomposition as an indicator and compare streams with different buffer treatments.
Knowledge about what variables affect leaf litter decomposition can help managers to take decisions that protect forest streams.
2
1.1 Aims of the study and scientific questions
The main objective of this study is to investigate the link between decomposition and riparian buffer management in Sweden. The aim is to address if the width of riparian buffers affects the rate of litter decomposition in small streams. As a complement to this, the relationships of litter decomposition to other variables known to be affected by riparian deforestation are addressed. The questions that the present thesis will attempt to answer are:
1. How is leaf litter decomposition in small streams affected by buffer width?
2. How does stream bottom habitat, light, stream velocity and water temperature affect decomposition?
2 Materials and methods
This thesis utilizes results that are part of a larger study at the Swedish University of Agricultural Sciences. The methods described here refer only to the parts of the study relevant to this thesis.
2.1 Study sites
The study was conducted in the counties of Västerbotten and Jönköping at twelve sites in both areas respectively. The sites in each region were situated within a 100 km radius (Fig.1).
They were selected to represent different riparian buffer widths and were divided into four categories: 1) no buffer, where no mature trees were left along the streams, 2) thin buffer, where 1-2 rows of trees were left along the streams (< 5 m wide), 3) moderate buffer which had trees > 5 m away from the steams and 4) reference sites (no harvest). In all other aspects, the streams were as similar to each other as possible. All sites, except reference streams, were placed in clear-cuts (with the particular buffers) which were harvested between 2010 and 2016. All streams were headwaters, ranging in catchment areas between 0.5 and 2.7 km2. For each stream, a 50m reach was selected as the study site. These reaches were situated at the most downstream part of the stream reach within the clear-cut. The buffer width along each of these reaches was measured at eight transects perpendicular to the stream channel, 4 transects at each side of the stream.
Fig 1. The northern sites marked on the left map and the southern sites marked on the right map. Colours of the
points represent the buffer treatment of ‘no buffer’ (red), ‘thin buffer’ (orange), ‘moderate buffer’ (blue) and
‘reference sites’ (green).
N N
3 2.2 Field sampling
The study was carried out from mid-September to mid-October 2018. Temperature was measured hourly over the whole period using two submerged HOBO® pendant loggers (Onset) at each site and then averaged per day (daily mean and over the whole period). Light was estimated as canopy openness measured at the beginning of the period using a spherical densitometer at each 2.5 m interval along the 50 m section. Percentages of different stream bottom substrates were assessed visually at ten quadrates (50x50 cm) at 5 m intervals along each 50 m stream section. The substrate was categorised as organic fraction, fine sediments (<0.25 – 2 mm), coarse sediments (2 – 400 mm) and large rocks/bedrock (> 400 mm). By measuring the time it took for a florescent fluid to travel a 3 m distance in the stream, current velocity was estimated once at the beginning of the period and once at the end.
2.3 Litter decomposition
To investigate leaf litter decomposition, birch litter was collected at a single location at abscission in the year preceding the study. The litter material was air-dried indoors at room temperature to a constant weight. 3.00 ± 0.01 g of litter material was inserted into coarse mesh bags with 5mm mesh to measure leaf litter decomposition in the presence of both macroinvertebrates and microbes. At each site, in mid-September, five litterbags were put in the water, at the stream bottom to an average stream depth, attached to a brick with a zip-tie.
The litter bags were collected in mid-October rendering a total of 34 – 36 field days. The litterbags were brought to the laboratory where the litter was removed from the bags and rinsed before being oven dried at 60°C for 48hr. The dried litter was then combusted at 550°C for 4hrs to obtain ash-free dry mass (AFDM) (Benfield 1996). Litter AFDM, used as a proxy for decomposition, was obtained by subtracting ash weight from dry mass. Following the same procedure, initial AFDM was obtained by combusting five 3g unconditioned samples. To correct for handling losses, five bags were also brought to the field and then combusted following the same procedure. The remaining AFDM was subtracted from the initial AFDM and then, as a proportion of the initial AFDM, used as a measure of
decomposition in the following analyses.
2.4 Statistical analyses
Analysis of variance (ANOVA) were used to analyse the effects of location (north or south) and buffer width category on litter decomposition. After first testing for, and finding a difference between northern and southern sites, all the consequent analyses were made for south and north separately. Because we expected decomposition being related to other than buffer width variables, we also tested the effects of all other environmental variables, which was done using linear regressions. The effects of buffer width as a continuous number was not preferred to be used in linear regression analysis with decomposition since there was no specific number of buffer width at the reference sites (i.e. they had infinite buffers) and they would thus have had to be excluded from the regressions. All data was averaged per site to make statistically independent data points. All analyses were performed in Microsoft Excel version 16.23 using the Analysis ToolPak.
4
3 Results
The southern sites showed significantly lower average values of remaining AFDM compared to the northern sites (p=2.52E-10, F=48.3) (Fig.2a). For both the northern and the southern sites there was no significant difference in litter decomposition between sites with different buffer widths (north: p=0.82, F=0.31 and south: p=0.37, F=0.08) (Fig.2b).
Fig.2. Mean values ± 1 SE of leaf litter decomposition as a function of location (a). Mean values ± 1 SE of leaf litter decomposition as a function of buffer width category in the north and in the south (b).
Mean values for all variables in this study is reported in Table 1. Many of the variables differ clearly between north and south (e.g. temperature, velocity and canopy openness) while some showed less apparent differences (e.g. stream bottom substrate). For most of the variables, the differences between buffer categories are small or almost negligible.
70 75 80 85
% Remaining AFDM
a
North South
70 75 80 85
No Thin Moderate Reference
% Remaining AFDM
Buffer width
b
North South
5
Table 1. Mean (µ), standard error, min- and max values for all variables.
Total North n=11, South N=12 Max 87.5 81.3 89.5 70.0 42.2 50.0 79.5 68.5 89.8 80.5 0.21 0.85 6.5 9.5
Min 78.0 70.4 5.0 9.0 0.0 2.0 0.0 0.0 8.0 5.3 0.02 0.04 5.3 8.6
SE 0.8 1.2 8.1 5.6 4.3 4.6 8.2 6.3 7.0 6.5 0.02 0.07 0.1 0.1
µ 82.4 76.1 31.6 36.9 18.7 23.7 44.8 19.8 18.3 38.5 0.10 0.27 5.9 9.0
Reference North n=3, South n=3 Max 84.6 80.0 10.0 70.0 35.5 27.0 66.0 42.0 16.3 13.9 0.07 0.28 6.5 9.4
Min 81.7 70.4 5.0 26.0 24.0 16.5 44.5 4.0 8.0 5.3 0.03 0.04 5.4 8.6
SE 0.9 2.8 1.7 12.7 3.3 8.9 17.6 12.3 2.7 2.5 0.01 0.07 0.3 0.2
µ 82.8 75.5 6.7 48.3 29.7 30.1 31.8 17.3 7.6 9.3 0.06 0.14 6.0 9.0
Moderate buffer North n=2, South n=4 Max 87.5 81.3 31.5 50.0 24.5 50.0 79.5 68.5 18.2 48.3 0.16 0.49 6.4 9.2
Min 78.0 71.9 11.0 9.0 12.5 12.0 0.0 0.0 8.9 12.1 0.03 0.07 5.5 8.7
SE 4.6 2.2 10.3 9.2 6.0 8.8 14.8 14.3 4.7 8.5 0.07 0.09 0.5 0.1
µ 82.7 78.3 21.3 36.0 18.5 24.0 58.8 20.3 13.5 19.3 0.09 0.30 5.9 9.0
Thin buffer North n=4, South n=2 Max 83.8 73.1 89.5 25.0 42.2 43.2 79.5 63.5 49.3 45.2 0.21 0.85 6.3 9.4
Min 79.3 73.0 13.0 14.3 0.0 2.0 0.0 27.5 36.4 30.3 0.07 0.19 5.9 8.8
SE 1.1 0.1 16.0 5.4 10.4 20.6 18.0 5.3 6.9 7.5 0.03 0.33 0.1 0.3
µ 82.4 73.0 50.2 19.7 11.2 22.6 19.3 41.3 47.5 37.7 0.12 0.52 6.1 9.1
No buffer North n=2, South n=3 Max 82.5 78.7 53.5 53.5 24.0 31.0 24.0 63.5 89.8 80.5 0.15 0.31 5.9 9.5
Min 80.2 70.9 30.0 30 11.0 6.0 11.0 27.5 78.8 76.2 0.13 0.12 5.3 8.6
SE 1.1 2.5 11.6 13.2 6.5 7.3 21.5 10.9 5.5 1.3 0.01 0.06 1.6 0.3
µ 81.4 75.8 41.6 38.0 17.5 17.5 37.5 42.3 84.3 32.6 0.14 0.20 5.6 9.0
% Remaining AFDM. North % Remaining AFDM. South % Organic substr. North % Organic substr. South % Fine substr. North % Fine substr. South % Coarse substr. North % Coarse substr. South % Open canopy North % Open canopy South Velocity (m/s) North Velocity (m/s) South Temp. (°C) North Temp. (°C) South
6
Regression analyses of the southern sites showed a significantly positive trend between remaining AFDM and % organic bottom substrate (p=0.003, Fig. 3a) and a negative significant trend for temperature (p=0.01, Fig. 3f). There was also a negative trend of remaining AFDM with % coarse bottom substrate in the south (Fig. 3c) which was however not significant (p=0.06). For the northern sites there was a positive trend between remaining AFDM and temperature (opposite to that of the southern sites, Fig. 3f), as well as a negative trend with velocity (Fig. 3e). Both these trends were however not significant (p=0.08 and p=0.07). Percent of fine bottom substrate (Fig. 3b) and % open canopy (Fig. 3d) showed no trends for either northern or southern sites.
Fig.3. Regression trends of leaf litter decomposition (remaining AFDM) in the northern (blue) and the southern
(orange) sites as a function of organic (a), fine (b) and coarse (c) bottom substrate, canopy openness (d), stream velocity (e) and water temperature (f) respectively. Values are averaged for each site. R squared, n, and p-values are shown above each regression plot.
65 70 75 80 85 90
0 20 40 60 80 100
% Remaining AFDM
% Organic bottom substrate
a
North South
65 70 75 80 85 90
0 20 40 60
% Remaining AFDM
% Fine bottom substrate
b
North South
65 70 75 80 85 90
0 20 40 60 80 100
% Remaining AFDM
% Coarse bottom substrate
c
North South
65 70 75 80 85 90
0 20 40 60 80 100
% Remaining AFDM
% Open canopy
d
North South
65 70 75 80 85 90
0,0 0,2 0,4 0,6 0,8 1,0
% Remaining AFDM
Velocity (m/s)
e
North South
65 70 75 80 85 90
5 6 7 8 9 10
% Remaining AFDM
Temperature (°C) f
North South
n = 11 R2 = 2.4E-06 F = 2.152E-05 p = 0.996
n = 12 R2 = 0.612 F = 15.75 p = 0.003
n = 11 R2 = 0.001 F = 0.01 p = 0.932
n = 11 R2 = 0.329 F = 4.42 p = 0.065
n = 12 R2 = 0.114 F = 1.28 p = 0.284 n = 12 R2 = 0.308 F = 4.46 p = 0.061
n = 12 R2 = 0.470 F = 8.88 p = 0.014 n = 12 R2 = 0.006 F = 0.06 p = 0.805
n = 12 R2 = 0.001 F = 0.01 p = 0.931
n = 11 R2 = 0.303 F = 3.91 p = 0.079 n = 11 R2 = 0.020 F = 0.18 p = 0.677
n = 11 R2 = 0.005 F = 0.04 p = 0.841
7
4 Discussion
One of the strongest results in this study is the clear difference in amount of decomposition between the northern and the southern sites with higher decomposition in the south. This is expected since the temperature was also generally higher in the south (Table 1) and
temperature is known to drive decomposition through increased microbial biomass and/or enzymatic activity at elevated temperatures (Fenoy et al. 2016, Friberg et al. 2009).
Buffer width was also expected to have a clear effect. In a review by Sweeney and Newbold from 2014 it is concluded that a buffer zone of ≥30m is needed to protect the physical, chemical and biological integrity of small streams. In this study buffers were much narrower than 30 m and thus we expected that they would not maintain all their ecological functions.
However, buffer width had no effect on decomposition, neither in the south nor in the north.
The reason for this might have been that other variables were more important at this stage of litter decomposition (initial decomposition phase).
Organic bottom substrate showed a strong positive trend with remaining AFDM in this study.
Many detritivorous insects are dependent on a habitat with a coarse bottom substrate (Allan 1995). If their habitat is covered by organic material, these might be less abundant, thus slowing down decomposition. The close to significant negative trend for remaining AFDM with % coarse bottom substrate could support this theory. Interestingly, I only saw these trends in the southern sites, while I saw no trend between decomposition and stream bottom substrate in the north. Allan (1995) describes how the first 5 – 25 % of the weight loss after leaves have entered the water is due to leaching of soluble components and not
decomposition by biological communities. The decomposition was generally slower in the north, compared to the south. Even though the leaf packs were in the water for around 35 days, northern sites still averaged 82.4 % remaining AFDM (Table 1). It is therefore possible that hardly any colonisation of the biological communities in the streams took place and all or most of the weight loss might then be due to leaching. We know that temperature is a strong driving factor for decomposition, and we see it in the general difference between north and south. The expected effect of temperature, with higher decomposition in increasing temperature, is visible in the south but not in the north. Possibly, the northern stream biota was more strictly tied to colder water and it might then have become too hot for them, leading to a reverse effect of lower decomposition with increasing temperature. The trend between AFDM and stream velocity in the north would, if it had been significant, have agreed with the physical influences on litter decomposition (leaching) rather than biological being important at so low levels of decomposition.
Canopy openness, which showed no effect on decomposition in this study, was the variable with strongest connection to birch leaf litter decomposition in the study by Lidman et al.
(2017). The litter bags in the Lidman paper were however in the water for a longer period (53 – 56 days) after which there was only 48 % remaining AFDM, suggesting this effect might only show up after longer conditioning.
There might be several reasons behind the somewhat surprising results in this study. It was conducted after one of the hottest and driest summers, and during one of the hottest and driest autumns, ever recorded in Sweden (SMHI 2018a and 2018b). With the previously established importance of riparian buffers for temperature control in the stream (Sweeney et al. 2014), buffer width should have been more important in these conditions and not
completely unimportant, as the results in this study shows. Of the streams investigated in this study, all of those in the southern part of Sweden and many of those in the northern part of Sweden dried out during the summer. The effects of drought on Swedish aquatic
ecosystems is largely unknown but results from Portugal shows that at least the microbial detritivore community and subsequently leaf litter decomposition is negatively impacted by drought (Mora-Gómez 2018). It is possible that the detritivorous inhabitants of the streams,
8
regardless of buffer width, did not survive such extreme conditions and there simply might not have been any, or enough, detritivores left to decompose the leaf litter during the autumn. It is also possible that the leaf packs would have needed to be in the water for a longer period to be able to indicate any discernible trends. With the slow decomposition, any differences between sites with different buffer width, canopy openness etc., might not have had time to show up yet in this study. Tolkkinen (2013) did show results after 30 days in similar environment as our but they used litter from alder. Different species do generally decompose very differently so it is possible this was enough time for alder but not for birch (Allan 1995, Kuglerová et al. 2017, Lidman et al. 2017).Finally, it is possible that more data points, obtained via a more rigorous statistical analysis using multiple samples from each site, could have yielded significant results for those trends (coarse bottom substrate in the south and velocity and temperature in the north) that with the methods used here were promisingly strong but did not in the end show significance.
In conclusion, although organic bottom substrate and temperature (in the southern sites) were the only variables with significant effects on leaf litter decomposition and buffer width did not matter, it is impossible to determine if this lack of significant differences is due to the drought, the statistical method or that there actually are no real differences. It would be interesting to revisit these sites during a year with more average weather to compare the results. Despite having few clear results, the trends with organic bottom substrate and temperature that we do see still point towards the importance of managers working to target particular properties of riparian function that protects ecological integrity, specifically temperature control and sediment retainment and especially along smaller streams. This might not then be directly connected to buffer width but rather to buffer function and more studies in this area is needed to determine how small forest streams should be managed in the light of hotter and dryer climate possibly becoming more common during Swedish summers.
9
References
Allan, J. David. 1995. Stream ecology, structure and function of running waters. London:
Chapman & Hall.
Clapcott, Joanne E., Collier, Kevin J., Death, Russel G., Goodwin, Eric O., Harding, Jonathan S., Kelly, Dave., Leathwick, John R. and Young, Roger G. 2012. Quantifying relationships between land-use gradients and structural and functional indicators of stream ecological integrity. Fresh water Biology 57(1): 74–90
Erdozain, Maitane., Kidd, Karen., Kreutzweiser, David and Sibley, Paul. 2018. Linking stream ecosystem integrity to catchment and reach conditions in an intensively managed forest landscape. Ecosphere 9(5): e02278. 10.1002/ecs2.2278
Fenoy, Encarnación., Casas, J. Jesús., Díaz-López, Manuel., Rubio, Juan., Guil-Guerrero, J.
Luís and Moyano-López, Francisco J. 2016. Temperature and substrate chemistry as major drivers of interregional variability of leaf microbial decomposition and cellulolytic activity in headwater streams. FEMS Microbiology Ecology 92(11): fiw169.
Friberg, Nikolai., Dybkjær, John B., Olafsson, Jon S., Gislason, Gisli Mar., Larsen Søren E.
and Lauridsen, Torben L. 2009. Relationships between structure and function in streamscontrasting in temperature. Freshwater Biology 54: 2051–2068.
Hynes, H. B. Noel. 1975. The stream and its valley. Verhandlungen Des Internationalen Verein Limnologie 19:1–15.
Kuglerová, Lenka., García, Liliana., Pardo, Isabel., Mottiar, Yaseen and Richardson, John. S.
2017. Does leaf litter from invasive plants contribute the same support of a stream ecosystem function as native vegetation? Ecosphere 8, e01779.
Lidman, Johan., Jonsson, Micael., Burrows, Ryan M., Bundschuh, Mirco and Sponseller, Ryan A. 2017. Composition of riparian litter input regulates organic matter decomposition:
Implications for headwater stream functioning in a managed forest landscape. Ecology and Evolution 7 (4): 1068–1077.
Mora-Gómez, Juanita., Duarte, Sofia., Cássio, Fernanda., Pascoal, Cláudia., and Romaní, Anna M. 2018. Microbial decomposition is highly sensitive to leaf litter emersion in a permanent temperate stream. Science of The Total Environment 621:496-496.
Naiman, Robert J. and Décamps, Henri., 1997. The ecology of interfaces: Riparian Zones.
Annual Review of Ecology and Systematics 28: 621–658.
Richardson, John S., Naiman, Robert J. and Bisson, Peter A. 2012. How did fixed-width buffers become standard practice for protecting freshwaters and their riparian areas from forest harvest practices? Freshwater Science 31, 232–238.
Swedish Meteorological and Hydrological Institute (SMHI). 2018a. Sommaren 2018 – extremt varm och solig. https://www.smhi.se/klimat/klimatet-da-och-nu/arets- vader/sommaren-2018-extremt-varm-och-solig-1.138134 (Hämtad 2019-04-24).
Swedish Meteorological and Hydrological Institute (SMHI). 2018a. Hösten 2018 – varm och torr. https://www.smhi.se/klimat/klimatet-da-och-nu/arets-vader/hosten-2018-varm- och-torr-1.139553 (Hämtad 2019-04-24).
Sweeney Bernard W., Bott, Thomas L., Jackson, John K., Kaplan, Louis A., Newbold, J.
Denis., Standley, Laurel J., Heisson, W. Cully and Horwitz, Richard J. 2004. Riparian deforestation, stream narrowing, and loss of stream ecosystem services. Proceedings of the National Academy of Sciences 101: 14132-14137.
Sweeney, Bernard W. and Newbold, J. Denis. 2014. Streamside forest buffer width needed to protect stream water quality, habitat, and organisms: a literature review. Journal of the American Water Resources Association 50:560-584.
Tolkkinen, Mikko., Mykrä, Heikki., Markkola, Anna-Mari., Aisala, Heidi., Vuori, Kari-Matti., Lumme, Jaakko., Pirttilä, Anna Maria and Muotka, Timo. 2013. Decomposer communities in human-impacted streams: species dominance rather than richness affects leaf
decomposition. Journal of Applied Ecology 50(5): 1142-1151.
Zinko, Ursula. 2005. Strandzoner längs skogsvattendrag. Världsnaturfonden WWF.
