Examensarbete i biologi avseende kandidatexamen 15 hp VT 2021
TESTING THE INFLUENCE OF RIPARIAN BUFFER DESIGN ON STREAM BIODIVERSITY
FOLLOWING DROUGHT
Rasmus Eriksson
En undersökning av kantzoners påverkan på den biologiska mångfalden av makroevertebrater efter en period av torka
Rasmus Eriksson
Abstract
Forestry is a major industry in Sweden and the most common method to harvest timber is to clear-cut large areas. Clear-cutting can alter multiple physical, chemical, and ecological characteristics of headwater streams. To minimize these effects, vegetated riparian ‘buffers zones’ are often spared along streams and lakes. Despite this, there are uncertainties regarding optimal width of buffers to safeguard streams from clearcutting effects. In this study, I ask how riparian buffer width influences stream macroinvertebrate communities, and how this influence may be altered by summer drought. I tested this in 24 headwater streams, half located in northern Sweden and the other half in southern Sweden. Streams in each region included four different buffer width categories (n = 3), including: “no buffer” (no trees left post-harvest), “thin buffer” (< 5 m), “moderate buffer” (> 5 m) and “reference” (no harvest). I analysed a suite of metrics that describe the abundance, richness, and composition of macroinvertebrates, and compared these across streams with different buffer properties.
Regionally, southern streams had marginally greater taxonomic richness and relative abundance of sensitive taxa compared to northern counterparts, regardless of buffer conditions. Further, thin and absent buffers performed the best across several
macroinvertebrate metrics, particularly for southern streams. Antecedent drought had no observable effects on macroinvertebrate communities, but taxonomic richness across region was positively correlated with stream pH. Overall, my findings, while tentative given low statistical power, suggest that retaining coniferous-dominated buffers may not lead to the desired ecological outcomes in boreal headwaters.
Key words: forestry, clear-cutting, riparian buffers, macroinvertebrates, drought.
Table of contents
1. Introduction ... 1
1.1 Aim of study and scientific questions ... 2
2. Method ... 3
2.1 Study sites ... 3
2.2 Field sampling ... 4
2.3 Statistical analyses ... 4
3. Results ... 5
4. Discussion ... 9
5.1 Buffer zone effects ... 9
5.2 Drought effects ... 10
5.3 Regional effects ... 11
5.4 Methodological considerations ... 11
5.5 Conclusions and the future ... 12
Acknowledgements ... 12
References ... 13
1
1. Introduction
Anthropogenic pollution occurs all throughout the world and it negatively impacts both terrestrial and aquatic ecosystems alike (Owens, Williams and Haynes 2021). Aquatic ecosystems are threatened and impacted by land use such as agriculture, urbanization and forestry (Owens, Williams and Haynes 2021). In unison, these forms of land use can lead to degradation of aquatic ecosystems in the form of reduced biodiversity and weakened ecosystem services (Forio et al. 2020). There are several measures that can be applied to stabilize degraded aquatic systems such as streams (Forio et al. 2020). Removal of dams, construction of fish passages, restrictions regarding fertilizer use and conservation of riparian vegetation are examples of such measures (Forio et al. 2020).
Headwater streams, the smallest streams of a stream network, account for upwards of 80 % of the total network length and function as important habitats for biological diversity in the form of macroinvertebrates (Jyväsjärvi, Koivunen and Muotka 2020). Smaller streams are also known to contain unique organisms that are seldom found in larger streams (Kuglerová et al. 2020). Headwater streams proximity to riparian vegetation causes them to be
vulnerable to forestry practices (Kuglerová et al. 2020). Despite this, protecting headwater streams is often overlooked by forestry companies (Jyväsjärvi, Koivunen and Muotka 2020).
Riparian buffers consist of a strip of riparian vegetation that serves to protect the health of the stream by reducing nutrient and sediment runoff from land, stabilizing stream banks with their roots, maintaining water temperature through shading and providing shelter for macroinvertebrates that require terrestrial resources (Jyväsjärvi, Koivunen and Muotka 2020). Riparian buffers are also believed to act as important habitats and corridors for the movement and dispersal of some macroinvertebrate species (Kuglerová et al. 2020).
Furthermore, forest debris from riparian buffers can impact the stream morphology and flow by trapping sediment and organic material which slows the stream down as well as creating habitats for macroinvertebrates (Grimstead, Krynak and Yates 2018).
Forest management in Fennoscandia mainly consists of clear-cutting, where all trees regardless of size and species are cut down within a certain plot of land (Lundmark,
Josefsson and Östlund 2013). The area then remains deforested for a period of time before being resown (Lundmark, Josefsson and Östlund 2013). Clear-cutting can affect the
microclimate of the area by altering the water cycle, soil properties and the species diversity both on land and in streams (Keenan and Kimmins 1993). Although the extent of the effects depends on the climate, topography and geology of the site that is to be deforested (Keenan and Kimmins 1993). A comparison of riparian buffer practices among three major forestry countries in the northern hemisphere consisting of Canada, Sweden and Finland, revealed that the average width of riparian buffers around headwater streams was smaller than recommendations set by most scientific studies (Jyväsjärvi, Koivunen and Muotka 2020).
Several studies indicate that riparian buffers must be at least 30 meter wide to protect the environmental conditions, ecosystems services and biodiversity of streams (Jyväsjärvi, Koivunen and Muotka 2020). Regardless of this, riparian buffers are only recommended to be at least 5 meter wide in Finland and 10 meter wide in Sweden according to the
Endorsement of Forest Certification (PEFC) (Jyväsjärvi, Koivunen and Muotka 2020;
Skogforsk 2018).
Benthic macroinvertebrate communities can be used as a measure of water quality and the
overall biological integrity of an aquatic ecosystem through using various indices (Owens,
Williams and Haynes. 2020). Macroinvertebrates are excellent biological indicators because
2
they are mainly positioned in the middle of the food chain, are affected by long- and short- term stress and are very diverse and abundant as a community (Owens, Williams and Haynes 2021). They are also relatively immobile which makes them reflect the local condition of streams while also being easy to collect (Etemi et al. 2020). Due to the large diversity of macroinvertebrates, there are both species that are highly tolerant and species that are highly sensitive to different forms of disturbances (Owens, Williams and Haynes 2021). The EPT index is a macroinvertebrate index that is commonly used to measure water quality (WSI n.d.). It involves three macroinvertebrate orders in Ephemeroptera (mayflies), Plecoptera (stoneflies) and Trichoptera (caddisflies) that are all sensitive to organic pollution (WSI n.d.).
Macroinvertebrates can also be roughly divided into functional feeding groups depending on how they acquire nutrition (MPWMD 2004). These include scrapers/grazers that consume algae, shredders that consume organic litter such as leaves, collector-gatherers that collect material from the bottom of the stream, collector-filterers that filter particles from the water column, and finally predators that feed on other macroinvertebrates (MPWMD 2004).
Disturbances from clear-cutting could lead to changes in the macroinvertebrate community where a reduction in shredder species is followed by an increase in grazers or collector- gatherers (Kreutzweiser, Capell and Good 2005).
The increase in global average temperatures due to climate change is causing droughts to become more common in boreal regions, which in turn could lead to severe changes within stream macroinvertebrate communities (Spinoni et al. 2018). Droughts cause increased fragmentation and habitat loss that can reduce the biodiversity and function of the ecosystem and, in severe cases, cause a collapse of the entire stream food web (Lu et al. 2016). The ability for a macroinvertebrate community to recover depends on the recolonization abilities of its species (Lake 2003). Unfortunately, not much is known about macroinvertebrate community responses to drought in boreal streams since droughts are still relatively uncommon events (Sarremejane et al. 2020). Drying could lead to a more homogenous community if it only contains a few drought tolerant species, or a more heterogenous community if there are many species that carry drought tolerant traits (Sarremejane et al.
2020).
For this thesis I will study the effect of riparian buffer widths on macroinvertebrate community composition in northern and southern Sweden after a prolonged period of drought in the summer of 2018. That summer was breaking records with the highest temperatures ever measured in certain parts of Sweden (SMHI 2018). Through using macroinvertebrate and temperature data from sites in both regions, I will be able to determine if there is any difference in resilience between stream macroinvertebrate communities as a response to drought.
1.1 Aim of the study and scientific questions
The primary goal of this thesis project is to test the effects of riparian buffer width on stream macroinvertebrate communities in northern and southern Sweden following a period of prolonged drought during the summer of 2018. The scientific questions that I will attempt to answer in this thesis are the following:
1. How does riparian buffers with varying widths effect stream macroinvertebrate communities?
2. What impact does a prolonged period of drought have on stream macroinvertebrate
communities?
3
2. Method
The data used in this thesis were collected and identified from late June to October 2018 as a part of the SOSTPRO project, which stands for Source Stream (headwater streams)
Protection from forest practices. The project is a collaboration between University of British Columbia in Canada, University of Oulu in Finland and the Swedish Agricultural University in Sweden. Lenka Kuglerová, the leader for the Swedish side of the SOSTPRO project, provided me with the macroinvertebrate and temperature data that I use for my analysis in this thesis.
2.1 Study sites
The macroinvertebrate and temperature data were collected from 24 headwater streams in total, 12 in Jönköping County in southern Sweden and 12 in Västerbotten County in northern Sweden (Fig. 1). The streams were within a 100 km radius of one another in each county and the catchment sizes varied between 0,18-6,1 km
2and from 63-320 m above sea level in elevation (Table 1). The streams were divided into four categories of no buffer, thin buffer, moderate buffer and reference. The amount of riparian vegetation varied depending on the category, where no buffer meant that there were no mature trees, thin buffer meant 1-2 rows of trees (< 5 m wide), moderate buffer meant a > 5 m wide row of trees and the reference sites meant no harvest at all. All sites, excluding the reference sites, had been clear-cut somewhere between 2010 and 2016. Each study site consisted of a 50 m long reach and the riparian buffer width was measured four times, twice on each side of the stream, by using a total of eight transects that ran perpendicular to the stream itself. Each study site was placed at the most downstream area of the clear-cut.
Fig. 1: The northern sites in Västerbotten County (left) and the southern sites in Jönköping County (right). The sites are colour-coded according to buffer width category. Source: Edith Bremer.
In addition to buffer features, study sites also varied in terms of other chemical properties
(Table 1). For example, pH ranged from 4,3-7,3 across all sites (Table 1). Similarly, dissolved
organic carbon (DOC) had an almost sevenfold difference in concentration, ranging from 4,9-
34,5 mg/L among sites (Table 1). Total nitrogen correlated closely with DOC, apart from one
nitrogen rich site (thin buffer 3S), ranging from 0,15-2,5 mg/L (Table 1). Finally, phosphate
(PO4) varied from 1,0-7,7 µg/L between sites (Table 1).
4
Table 1: Water chemistry data from the study sites with different buffer categories, including moderate buffer widths (MB), thin buffers (TB), no buffers (NB), and reference sites (R), in the north (N) and south (S) of Sweden.
Values are averages from two sampling occasions in September and October. Note that one site (NB 1N) has no water chemistry data due to being destroyed by ditching.
Site Catchment
area (km2)
Elevation
(m) pH DOC
(mg/L)
Total N (mg/L)
PO4 (µg/L)
MB 1N 0,68 124 5,41 4,90 0,17 1,0
MB 2N 1,33 214 5,30 27,70 0,44 7,7
MB 3N 0,66 281 6,24 19,21 0,41 4,6
NB 1N nd nd nd nd nd nd
NB 2N 1,04 90 5,50 32,51 0,63 2,6
NB 3N 0,23 104 6,76 16,87 0,40 4,5
R 1N 1,13 81 6,57 10,78 0,30 3,2
R 2N 0,18 217 6,83 7,45 0,15 2,9
R 3N 0,52 211 5,99 16,36 0,32 3,6
TB 1N 2,33 63 6,72 12,89 0,29 5,1
TB 2N 1 235 5,85 15,75 0,28 2,5
TB 3N 0,9 136 5,96 16,12 0,38 4,8
MB 1S 1,17 207 6,12 24,62 0,48 3,7
MB 2S 3,95 234 6,49 34,50 0,78 1,6
MB 3S 1,4 172 4,91 18,72 0,55 2,0
NB 1S 1,33 229 6,87 11,43 0,43 3,5
NB 2S 1,87 244 6,80 24,65 0,63 3,3
NB 3S 1,93 222 5,77 21,28 0,56 1,4
R 1S 1,3 276 5,20 13,29 0,33 1,2
R 1S 2,73 219 6,45 19,20 0,68 3,5
R 3S 0,5 320 4,34 12,35 0,39 1,8
TB 1S 1,03 313 7,00 26,99 0,60 2,3
TB 2S 6,06 178 6,63 21,35 0,55 2,3
TB 3S 2,66 196 7,25 9,095 2,525 4,17
2.2 Field sampling
The macroinvertebrates were collected between mid-September to mid-October 2018 by using a Surber sampler with an area of 20x25 cm. Additionally, pH was measured twice per study site, once in September and once in October, by using a titrator. Temperatures were measured every hour from late June to mid-October using two submerged HOBO® pendant loggers that were randomly placed within 10 m from the beginning and end of the 50 m long stream reach.
2.3 Statistical analyses
All the temperature and macroinvertebrate data were analysed in Microsoft Excel. I used the temperature data to create scatter plots which I used to determine the number of dry days at each study site through comparing the air temperature with the water temperature. I
compiled the already identified macroinvertebrate samples by calculating total abundance, richness, EPT richness, percentage EPT, percentage Chironomidae and percentage
Ephemeroptera for each of the 24 study sites. I chose these metrics since they provide various kinds of information about the macroinvertebrate communities at each site. Firstly,
abundance and richness are general metrics that describe the capacity of the streams to support both productive and diverse assemblages. On the contrary, EPT richness and
percentage EPT abundance focus on the presence and relative abundance of orders that tend
to be more vulnerable to pollution in aquatic ecosystems. Percentage Chironomidae and
percentage Ephemeroptera abundance are indicator groups that may provide insight into the
drivers behind assemblage structure. Chironomids are interesting since they are a family of
Dipterans that can be highly abundant and dominate macroinvertebrate communities in
streams with poor water quality due to their high resilience and recovery rate (Miller and
Golladay 1996). Conversely, Ephemeroptera can be a useful indicator taxa to look at within
5
Sweden since they are highly sensitive to low pH, even when it is naturally low as in many northern headwater streams (Petrin, Laudon and Malmqvist 2007). I used these metrics to compare average values across each buffer category (reference, thin, moderate and none) and between the two regions using a two-way ANOVA with a p-value threshold of < 0,05.
However, p-values < 0,1 were also considered to be moderately significant given the low statistical power of this test due to a low number of study sites. I used correlation analyses to test for any relationships between invertebrate metrics and water chemistry variables, including pH, DOC, total nitrogen and phosphate. Finally, I also used correlation analyses to determine if the macroinvertebrate metrics were at all correlated with the number of dry days during the summer of 2018.
3. Results
The mean abundance of macroinvertebrates varied among sites as well as across buffer categories. Mean abundance ranged from 1873,3-4369,3 individuals/m
2in the northern sites and 1932,3-2801,3 individuals/m
2in the southern sites (Fig. 2). The site with the highest abundance in the north had a thin riparian buffer, whereas the southern site with the highest abundance had no buffer at all (Fig. 2). However, difference in abundance across region or buffer were not statistically significant (p > 0,05; Appendix, Table 1A).
Fig. 2: Mean macroinvertebrate abundance at both sites (north & south) and their buffer categories. Error bars show the standard error.
Macroinvertebrate community richness also varied across sites and buffer categories, but the effects of buffer type was not statistically significant at the 0,05 level (Appendix, Table 1A).
The effect of region was, however, marginally significant (p = 0,058; Appendix, Table 1A), indicating greater overall richness in the southern compared to northern streams (Fig. 3a). In general, richness and EPT richness had similar trends, such that streams with thin buffers had both the highest EPT richness with an average of 13,3 species (SE = 1,3) and the highest overall richness for the southern sites with 35 species and families (SE = 3,6) (Fig. 3ab). Yet, only for EPT richness was there a marginally significant interactive effect between region and buffer category (p = 0,085; Appendix, Table 1A).
0 1000 2000 3000 4000 5000 6000 7000
Reference Moderate buffer Thin buffer No buffer
Abundance/m2
Northern sites Southern sites
6
Exploring this more deeply, overall richness and EPT richness were notably similar for northern sites across all buffer categories. By comparison, in the south, the reference sites had a noticeably lower richness with an average of 17,6 (SE = 2,0) when compared to the other categories (Fig. 3a). The reference sites in the south also had the lowest EPT richness with an average of 5 species (SE = 1,0) (Fig. 3b).
Fig. 3a: Overall mean macroinvertebrate community richness at both sites (north & south) and their buffer categories. Fig. 3b: Mean EPT richness at both sites and their buffer categories. Error bars show the standard error.
The EPT relative (%) abundance varied between northern and southern streams, yet this pattern was only moderately significant (p = 0,072; Appendix, Table 1A). As with richness and EPT richness, average % EPT was not significantly different across buffer categories (p >
0,05). In the northern region, sites with moderate buffers tended to have greatest average EPT relative abundance, where a total of 32,2 % of all macroinvertebrates consisted of taxa within these groups (Fig. 4a). By comparison, the greatest average EPT relative abundance for the southern sites was found in the reference sites, where EPT species made up 43,7 % (SE = 16,3) of all macroinvertebrates (Fig. 4a). This is in contrast with the reference sites in the north that instead had the lowest average EPT abundance at 13,5 % (SE = 4,9) (Fig. 4a).
The lowest average EPT abundance across the southern sites was shared between sites with thin buffers and no buffers at 31,9 % (SE = 6,9, 6,1) (Fig. 4a).
The average % abundance of Chironomidae differed by region (p = 0,027; Fig 4b) and tended to be higher in the northern sites across all buffer categories except for no buffer when compared to the southern sites. Furthermore, there was a marginally significant interactive effect of region and buffer category on % Chironomidae (p = 0,055; Appendix, Table 1A).
Accordingly, the greatest abundance of Chironomidae in the northern sites was found in sites with thin buffers where they made up 68,1 % (SE = 15,2) of all macroinvertebrates (Fig. 4b).
The southern sites had the greatest average Chironomidae abundance, 40,8 % (SE = 16,7), in sites with no buffer at all (Fig. 4b).
0 5 10 15 20 25 30 35 40 45
Reference M buffer T buffer No buffer
No. of species/families
Northern sites Southern sites
0 2 4 6 8 10 12 14 16 18
Reference M buffer T buffer No buffer
No. of EPT species
Northern sites Southern sites
7
Fig. 4a: Mean EPT abundance in percentage at both sites (north & south) and their buffer categories. Fig. 4b:
Mean Chironomidae abundance in percentage at both sites (north & south) and their buffer categories. Error bars show the standard error.
The Ephemeroptera relative abundance ranged between 0,6 – 7,8 % in the northern sites and 0,1 – 7,8 % in the southern sites (Fig. 5). In the north, Ephemeroptera (mayflies) abundance was greatest at sites with no buffer where they made up 7,8 % (SE = 7,8) of all
macroinvertebrates (Fig. 5). By contrast, the greatest average % Ephemeroptera abundance in the south was found at sites with thin buffers, also making up 7,8 % (SE = 3,0) of the total (Fig. 5). The Ephemeroptera % abundance data had to be log transformed to meet the assumptions of the two-way ANOVA. This test suggested marginally greater % abundance of Ephemeroptera in southern compared to northern streams (p = 0,072; Appendix, Table 1A).
A non-parametric test using the untransformed data to compare % Ephemeroptera between the northern and southern region indicated a similar pattern with slightly stronger statistical confidence (p = 0,04). Finally, the two-way ANOVA results also indicated a moderately significant p-value of 0,091 regarding the effect from the different buffer categories,
apparently linked to elevated % Ephemeroptera in streams with thin or no buffers (Appendix, Table 1A).
Fig. 5: Mean Ephemeroptera abundance in percentage at both sites (north & south) and their buffer categories.
Error bars show the standard error.
0 2 4 6 8 10 12 14 16 18
Reference Moderate buffer Thin buffer No buffer
Ephemeroptera abundance (%)
Northern sites Southern sites 0
10 20 30 40 50 60 70
Reference M buffer T buffer No buffer
EPT abundance (%)
Northern sites Southern sites
0 10 20 30 40 50 60 70 80 90
Reference M buffer T buffer No buffer
Chironomidae abundance (%)
Northern sites Southern sites
8
In general, pH emerged as the only chemical variable that was clearly correlated with
variation in macroinvertebrate metrics. For example, both richness (r = 0,70, p = 0,001) and EPT richness (r = 0,67, p = 0,001) increased significantly as sites become more
circumneutral (pH = 6,5-7,5, Fig. 6ab). Similarly, % Ephemeroptera abundance tended to increase with greater pH, but this relationship was non-linear, and suggested a threshold pH value around 5,5, below which the group was largely absent (data not shown). No other chemical variables that were significantly related to variation in macroinvertebrate metrics.
Fig. 6a: Correlation between taxonomic richness and pH across all sites (r = 0,70, p < 0,001). Fig. 6b: Correlation between EPT richness and pH across all sites (r = 0,67, p < 0,001)
The pattern of drought suggested that the southern sites, in particular those with no buffers were hit the hardest with an average drought period of 30,3 days (Fig. 7). Despite this, there appeared to be no significant correlation between drought and any of the macroinvertebrate metrics (p values for all correlations > 0,05). For example, the southern no buffer sites had the second highest overall richness out of all categories despite having experienced the longest period of drought (Fig. 3a).
Fig. 7: Average number of dry days at each site (north & south) for each buffer category. Error bars show the standard error.
0 5 10 15 20 25 30 35 40 45
0,00 2,00 4,00 6,00 8,00
Richness
pH
0 2 4 6 8 10 12 14 16 18 20
0,00 2,00 4,00 6,00 8,00
EPT Richness
pH
0 5 10 15 20 25 30 35 40 45 50
Reference Moderate Thin No buffer
Average no. of dry days
Northern sites Southern sites
9
5. Discussion
Riparian buffers are used globally to help reduce the negative consequences of land use on aquatic ecosystems, including the effects of forestry (Kuglerová et al. 2020). Nevertheless, the value of riparian buffers, as implemented in the Swedish forest landscape, is currently up for debate (Hasselquist et al. 2021). Here, I assessed the effects of variable riparian buffer width on stream communities using several macroinvertebrate metrics. Interestingly, while
tentative given low statistical power, my results indicate that thin buffer zones – and in some cases no buffer zones – performed best across several metrics. At any rate, there was no clear evidence that the widest buffer, or even the reference streams, supported more biologically rich communities. Altogether, these results suggest that we may need to revisit how we approach and construct riparian buffer zones in the boreal landscape.
5.1 Buffer zone effects
As previously mentioned, the generally assumed positive effects of riparian buffer zones are currently being challenged in the literature. For example, there have been studies that show that riparian buffers are not sufficient to protect the biodiversity and functionality in highly polluted streams (Effert-Fanta, Fischer and Wahl 2019). One reason buffers may fail is that they are not designed in the correct way to achieve desired water quality and/or ecological outcomes. For example, the width of current riparian buffers may be inadequate since Kuglerová et al. (2020) found that most of their surveyed streams in Sweden had widths smaller than 5 meters. Indeed, other studies have shown that a 15-meter buffer width is necessary to protect the stream conditions, whereas a 30-meter buffer width may be required to positively affect the biodiversity of the stream (Forio et al. 2020; Jyväsjärvi, Koivunen and Muotka 2020). For the streams studied here, I found that the overall abundance of
macroinvertebrates was highest in sites having thin or no buffers in both northern and southern streams. However, abundance by itself is not a great metric to determine how well an aquatic ecosystem is faring. The reason for this is that a stream could have a large
abundance whilst being homogenous and home to only a select number of species. Yet, in the current study, higher abundance does not seem to indicate a more homogenous community as thin and no buffer sites tended to also have the highest overall taxonomic richness.
The idea that narrower buffers could have positive effects on abundance is consistent with a study by Effert-Fanta, Fischer and Wahl (2019), who found that both macroinvertebrate and fish abundance reacted positively to increased levels of watershed agriculture when paired with thinner riparian buffers. They argued that narrow riparian buffers resulted in reduced canopy cover, which in turn increased the periphyton biomass, which is a mixture of algae, cyanobacteria, microbes and detritus that serve as food for invertebrates (Effert-Fanta, Fischer and Wahl 2019). Similarly, Kiffney, Richardson and Bull (2004) observed greater abundance and biomass of Chironomids and Ephemeroptera with narrower buffer widths.
Consistent with these studies, my results showed that buffer categories had a moderately significant effect on Ephemeroptera, with sites having thin or absent buffers exhibiting the greatest Ephemeroptera % abundance. This result could be explained by increased algae growth due to elevated light and nutrient levels since many Ephemeroptera species are grazers that consume algae (Poff et al. 2006). Collectively, these findings indicate that
narrower, or entirely absent buffers could promote the abundance of some macroinvertebrate
groups. Importantly, increased light and nutrition may not be all good though, since too
much of both at the same time can cause water quality problems, including blooms of algal
species that are not as edible for many macroinvertebrates (Effert-Fanta, Fischer and Wahl
10
2019). However, as small forests streams in the Swedish landscape rarely have high concentrations of inorganic nutrients, the risk that more open canopies will result in such eutrophic conditions seems low (Burrows et al. 2021).
One major challenge with interpreting the influence of buffer width in this study is that the reference sites may not be that great in terms of water and habitat quality. The reason for this is that the forested reference streams primarily drain mature production forests with a long history of management (Kuglerová, personal communication). The reference sites forests are thus likely homogenous with most of the trees being Norway spruce or Scots pine, potentially with limited understory. Furthermore, the riparian forests around the reference sites have not been felled in a long time and therefore have not experienced much natural or
anthropogenic disturbance, which could actually improve buffer functioning (Jyväsjärvi, Koivunen and Muotka 2020). Overall, to test the effects of buffer width in the face of forest clear cutting, it would be better to have reference sites that are as unmanaged as possible.
However, it is a huge challenge to find reference streams without some history of land use, especially in southern Sweden where anthropogenic impacts are more widespread due to a greater population density and a longer history of land management.
5.2 Drought effects
The drought from the record warm summer of 2018 strangely enough did not appear to impact macroinvertebrate communities in the autumn, as there were no significant correlations between number of drought days and any of the macroinvertebrate metrics.
Studies have shown that drought periods as short as one-week can reduce macroinvertebrate taxonomic richness and abundance in streams (Sarremejane et al. 2020). Further, in this same study, three weeks flow resumption were insufficient for the macroinvertebrate community to make a complete recovery from the negative impacts of the drought although this recovery was facilitated by greater connectivity to other streams (Sarremejane et al.
2020). The sampling of macroinvertebrates for this study was performed in September and October, which is two to three months after the main heatwave in summer. Therefore, the macroinvertebrate communities in the streams might have had time to recover from the negative effects of the drought before the sampling occurred. Although, the recovery rate should be relatively slow since the studied streams were headwater streams that are small and rather isolated when compared to larger watercourses such as rivers. However, the method used in this study may not have been the best at capturing the severity and potential impacts of drought. There were only two temperature loggers at each study site, one at the start and one at the end of each 50-meter reach. Therefore, it is hard to determine the severity of drought at each site as there is some possibility that surface water could be retained upstream, downstream, or even in between these loggers. In other words, it is not possible to tell with certainty that the streams were completely or only partially dry. A
completely dry stream may have more severe implications than a partly dried stream that has some remaining water that can function as refuge for invertebrates (Lake 2003). One
solution to this could involve using a greater number of temperature loggers placed in closer intervals of each other. Another solution could be to modify the loggers into water
conductivity sensors since they are potentially better for capturing zero-flow conditions
(Paillex et al. 2019). Finally, the number of variables in this study could also have impacted
my ability to detect the effects of drought. For example, it is hard to determine if the result is
explained by a difference in number of dry days, buffer width, or a mix of the two.
11 5.3 Regional effects
The abundance of macroinvertebrates was not affected by region nor buffer categories, but region had a moderately significant effect on taxonomic richness and a significant effect on Chironomidae % abundance, as well as a moderately significant effect on EPT and
Ephemeroptera % abundance. One potential explanation for this regional pattern could be lower stream acidity in southern sites compared to northern ones. Indeed, total richness and EPT richness both increased across sites with pH, as observed elsewhere in the region (Petrin, Laudon and Malmqvist 2007). Similarly, my results highlight the particular
sensitivity of Ephemeroptera to pH, indicating a potential threshold value of approximately 5,5, below which sites did not have any taxa from this Order. Overall, it is difficult to know what is driving the variation in pH across streams. For example, relatively lower pH in the north of Sweden could be related to the spring flood that transports up to half of the annual runoff water from snowmelt in just three to four weeks (Laudon 2000). Snowmelt supplies large amounts of organic acids along with some anthropogenic acids (e.g. sulphur) that, after melting, enter streams and create an acid shock that significantly lowers the pH (Laudon 2000). However, DOC concentrations and pH were not correlated, suggesting that other sources of acidity or buffering are operating across the sites. Differences in geology (e.g., limestone vs. granite parent material) across northern and southern sites could also explain differences in pH levels and taxonomic richness. However, it is hard to say for certain that this is a causal factor since there can be local variations in limestone deposits even within the same region. Further, studies have shown that southern streams can have strong or weak pH buffering capabilities (Ågren and Löfgren 2012), and in my study both the highest and lowest pH was observed in the southern region. There is also a larger number of industries in
southern Sweden that deposit mineral acids which could lower the pH of streams (Petrin, Laudon and Malmqvist 2007). Regardless of the underlying cause, my results highlight pH as an important driver of regional patterns of macroinvertebrate richness that operates
regardless of local buffer characteristics.
Despite the relationships with pH, differences in climate, including air temperature, could be another factor that is responsible for regional differences in richness. For example,
differences in air temperature that are mirrored in mean water temperatures and growing season length may result in greater periphyton biomass and in turn a greater abundance of Ephemeroptera in southern compared to northern streams (Kiffney, Richardson and Bull 2004). The same relationship is true for narrow buffer widths, since their smaller canopy cover results in more light hitting the stream and raising the water temperature (Kiffney, Richardson and Bull 2004).
5.4 Methodological considerations
Results from this analysis need to be taken with some caution because the statistical power of the tests is low due to the low number of study sites. A greater number of sites in both regions would likely yield more robust results, providing a better overall test of how buffer width influences streams. In addition, increasing the number of sampling occasions throughout the year could potentially improve the study. First, this may result in a better overall description of local richness, as not all aquatic insect species co-occur at the same time. In addition, sampling earlier on, during the summer, may have provided a better test of drought effects.
However, greater sampling intensity would require time and money, both of which are
limiting factors when conducting scientific research. There is also an argument to be had
regarding the use of a categorical versus a continuous approach when considering buffer
width. Finding appropriate streams with similar buffer widths is difficult and was
12
accomplished as well as it could be. However, it is impossible to have streams with identical hydrological and chemical conditions, with the only difference being variable buffer widths.
One would like the sites to be as similar as possible in every regard except for buffer width when performing the ANOVA, thereby minimizing the variance within groups. An alternative would be to avoid groups of buffer categories and instead use a regression approach that treat buffer width as a continuous and independent variable. Yet, the challenge for this is that there are no buffer width measurements taken for the reference sites. Instead, they would have to be given a set value of, for example, 100 meters to be able to perform regression analyses. Therefore, the categorical approach felt like the better option.
5.5 Conclusions and the future
My results challenge the current view of riparian buffers mainly being valued for their width.
This raises the question: is riparian buffer width really the be-all and end-all, or are there other factors to consider? As of now, most buffers mainly consist of conifer trees in the form of the Norway spruce or Scots pine (Hasselquist et al. 2021). A literature review by
Hasselquist et al. (2021) suggests that there are ecological benefits to be had from diversifying riparian buffer zones to encompass more broadleaved species of trees. More broadleaves would result in higher quality leaf litter that could be consumed by a greater variety of macroinvertebrates (Hasselquist et al. 2021). The biodiversity would not only be improved within the stream, but also on land since broadleaves forests are more species-rich than their coniferous counterpart (Hasselquist et al. 2021). Similarly, partial harvesting with up to 50 % removal of buffers zones greater than 30 meters wide has shown to create very minor risks to aquatic macroinvertebrate communities (Kreutzweiser et al. 2010). Instead, partial harvesting may be able to cause an increase in aquatic biodiversity by creating a more heterogenous forest structure, which leads to a more dynamic light regime and greater amount and diversity of detritus supplied to streams (Jyväsjärvi, Koivunen and Muotka 2020). I think that the idea of partial harvesting is sound since it allows for the planting of broadleaves and diversification of riparian buffers. Although it might be hard to change it from concept into reality since implementation of forestry management practices are usually only done for mature production forests at the time of final felling (Hasselquist et al. 2021).
The idea of an optimal buffer width regardless of the location and climate of the stream seems faulty. This idea is based on the assumption that streams are identical to one another, whereas the reality is more complicated with differences in biodiversity, hydrology and biogeochemistry (Kuglerová et al. 2017). The diversity of water conditions within different headwater streams calls for diversity in buffers. This study, with its low power, is not
providing enough evidence for us to be able to dispel the current notion of 15–30-meter-wide buffers being best suited to preserve biodiversity in streams. However, it is apparent that we still do not fully grasp the positives and negatives of different buffer widths and that further studies are needed to figure out the best course of action when it comes to the creation and management of riparian buffers in the future.
Acknowledgements
I would like to thank both Ryan Sponseller and Darshanaa Chellaiah for all their help in the
creation of this thesis. Ryan for his supervision and help with creating this thesis from start
to finish and Darshanaa for her guidance regarding the datasets. Lastly, I would also like to
thank Lenka Kuglerová for letting me use data that was collected as a part of the SOSTPRO
project.
13
References
Burrows, Ryan M., Jonsson, Micael., Fältström, Emma., Andersson, Jannika and Sponseller, Ryan A. 2021. Interactive effects of light and nutrients on stream algal growth modified by forest management in boreal landscapes. Forest Ecology and Management. 492 (2021) 119212
Effert-Fanta, Eden L., Fischer, Robert U. and Wahl, David H. 2019. Effects of riparian forest buffers and agricultural land use on macroinvertebrate and fish community structure.
Hydrobiologia. 841: 45-64
Etemi, Ferdije Z., Bytyçi, Pajtim., Fetoshi, Osman., Ymeri, Prespa., Albona, Shala-Abazi., Muja-Bajraktari, Nesade and Czikkely, Marton. 2020. The use of macroinvertebrate based biotic indices and diversity indices to evaluate the water quality of Lepenci river basin in Kosovo. Journal of Environmental Science and Health, Part A. 55 (6): 748-758
Forio, Marie A.E., De Troyer, Niels., Lock, Koen., Witing, Felix., Baert, Lotte., De Saeyer, Nancy., Rîsnoveanu, Geta., Popescu, Cristina., Burdon, Francis J., Kupilas, Benjamin., Friberg., Boets, Pieter., Volk, Martin., McKie, Brendan G. and Goethals, Peter. Small Patches of Riparian Woody Vegetation Enhance Biodiversity of Invertebrates. 2020. Water. 12: 3070 Grimstead, Jeremy P., Krynak, Edward M. and Yates, Adam G. 2018. Scale-specific land cover thresholds for conservation of stream invertebrate communities in agricultural landscapes. Landscape Ecology. 33: 2239-2252
Hasselquist, Eliza M., Kuglerová, Lenka., Sjögren, Jörgen., Hjältén, Joakim., Ring, Eva., Sponseller, Ryan A., Andersson, Elisabet., Lundström, Johanna., Mancheva, Irina., Nordin, Annika and Laudon, Hjalmar. 2021. Moving towards multi-layered, mixed-species forests in riparian buffers will enhance their long-term function in boreal landscapes. Forest Ecology and Management. 493 (2021) 119254
Jyväsjärvi, Jussi., Koivunen, Iina and Muotka, Timo. 2020. Does the buffer width matter:
Testing the effectiveness of forest certificates in the protection of headwater stream ecosystems. Forest Ecology and Management 478: 118532
Keenan, Rodney J. and Kimmins, Hamish, J.P. 1993. The ecological effects of clear-cutting.
Environmental Reviews. 1 (2): 121-144
Kiffney, Peter M., Richardson, John S. and Bull, Jennifer P. 2004. Establishing light as a causal mechanism structuring stream communities in response to experimental
manipulation of riparian buffer width. Journal of the North American Benthological Society.
23 (3): 542-555
Kreutzweiser, David P., Muto, Elisa., Holmes, Stephen and Gunn, John J. 2010. Effects of upland clearcutting and riparian partial harvesting on leaf pack breakdown and aquatic invertebrates in boreal forest streams. Freshwater Biology. 55 (11): 2238-2252
Kreutzweiser, David P., Capell, Scott S. and Good, Kevin P. 2005. Macroinvertebrate
community responses to selection logging in riparian and upland areas of headwater
catchments in a northern hardwood forest. Journal of the North American Benthological
Society 24 (1): 208-222
14
Kuglerová, Lenka., Jyväsjärvi, Jussi., Ruffing, Claire., Muotka, Timo., Jonsson, Anna., Andersson, Elisabet and Richardson, John S. 2020. Cutting edge: A Comparison of Contemporary Practices of Riparian Buffer Retention Around Small Streams in Canada, Finland and Sweden. Water Resource Research. 56 (9)
Kuglerová, Lenka., Hasselquist, Eliza M., Richardson, John S., Sponseller, Ryan A., Kreutzweiser, David P. and Laudon, Hjalmar. 2017. Management perspectives on Aqua incognita: Connectivity and cumulative effects of small natural and artificial streams in boreal forests. Hydrological Processes. 31 (23): 4238-4244
Lake, P. S. 2003. Ecological effects of perturbation by drought in flowing waters. Freshwater Biology. 48 (7): 1161-1172
Laudon, Hjalmar. 2000. Separating Natural Acidity from Anthropogenic Acidification in the Spring Flood of Northern Sweden. Doctor’s diss., Swedish University of Agricultural Sciences. ISSN 1401-6230, ISBN 91-576-5894-3
Lu, Xueke., Gray, Clare., Brown, Lee E., Ledger, Mark E. Milner, Alexander M., Mondragón, Raúl J., Woodward, Guy and Ma, Athen. 2016. Drought rewires the cores of food webs.
Nature Climate Change. 6: 875-878
Lundmark, Hanna., Josefsson, Torbjörn and Östlund, Lars. 2013. The history of clear-cutting in northern Sweden – Driving forces and myths in boreal silviculture. Forest Ecology and Management. 307: 112-122
Miller, Maria A. and Golladay, Stephen W. 1996. Effects of spates and drying on
macroinvertebrate assemblages of an intermittent and a perennial prairie stream. Journal of the North American Benthological Society. 15 (4): 670-684
Monterey Peninsula Water Management District (MPWMD. 2004. Analysis of macroinvertebrate functional groups.
https://www.mpwmd.net/programs/river/watershed_assessment/5_7/5_7_2a/5_7_2_a_t ext.pdf (Retrieved 2021-04-15)
Owens, Mitchell C., Williams, Clayton J. and Haynes, James M. 2021. Effects of forested buffers on benthic macroinvertebrate indicators of water quality in the Western Finger Lakes, New York. Inland Waters. 11 (1): 78-88
Paillex, Amael., Siebers, Andre R., Ebi, Christian., Mesman, Jorit and Robinson, Christopher T. 2019. High stream intermittency in an alpine fluvial network: Val Roseg, Switzerland.
LIMNOLOGY and OCEANOGRAPHY. 65 (3): 557-568
Petrin, Zlatko., Laudon, Hjalmar and Malmqvist, Björn. 2007. Does freshwater
macroinvertebrate diversity along a pH-gradient reflect adaptation to low pH? Freshwater biology. 52: 2172-2183
Poff, LeRoy N., Olden, Julian D., Vieira, Nicole K. M., Finn, Deb. 2006. Functional Trait Niches of North American Lotic Insects: Traits-Based Ecological Applications in Light of Phylogenetic Relationships. Journal of the North American Benthological Society. 25 (4):
730-755
Sarremejane, Romain., Truchy, Amélie., McKie, Brendan G., Mykrä, Heikki., Johnson,
Richard K., Huusko, Ari., Sponseller, Ryan A. and Muotka, Timo. 2020. Stochastic processes
15
and ecological connectivity drive stream invertebrate community responses to short-term drought. Journal of Animal Ecology. 90 (4): 886-898
Skogforsk. 2018. Good practices for forest buffers to improve surface water quality in the Baltic Sea region.
https://www.skogforsk.se/contentassets/c091f11901904a7e982b156fc39d2a62/arbetsrappor t-995-2018.pdf (Retrieved 2021-04-08)
Spinoni, Jonathan., Vogt, Jürgen V., Naumann, Gustavo., Barbosa, Paulo and Dosio, Alessandro. 2018. Will drought events become more frequent and severe in Europe?
International Journal of Climatology. 38 (4): 1718-1736
Sveriges Meteorologiska och Hydrologiska Institut (SMHI). 2018. 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 (Retrieved 2021-04-02)
Watershed Science Institute (WSI). n.d. The EPT Index – NRCS National Water and Climate Center.
https://www.wcc.nrcs.usda.gov/ftpref/wntsc/strmRest/wshedCondition/EPTIndex.pdf (Retrieved 2021-04-13)
Ågren, Anneli and Löfgren, Stefan. 2012. pH sensitivity of Swedish forest streams related to
catchment characteristics and geographical location – Implications for forest bioenergy
harvest and ash return. Forest Ecology and Management. 276 (2012): 10-23
1
Appendix – ANOVA results
Appendix; Table 1A: ANOVA-analysis between different macroinvertebrate categories, buffer category and region.
P-values < 0,05 are significant, but p-values < 0,1 are also considered to be moderately significant.
Categories D F
Abundance Richne ss
EPT richness
EPT abundance
Chironomidae abundance (%)
Ephemeroptera abundance (%)
Region 1 0,362 0,058 0,204 0,072 0,027 0,072
Buffer category
3 0,630 0,223 0,472 0,679 0,590 0,091
Region x buffer category
3 0,820 0,307 0,085 0,553 0,055 0,501
