OPCE-STR.vp Acta Bot. Croat. 68 (2), 367–380, 2009 CODEN: ABCRA 25 ISSN 0365–0588 Influence of environmental and spatial variables on the distribution of surface sediment diatoms in an upland loch, Scotland HONG YANG*, ROGER J. FLOWER, RICHARD W. BATTARBEE Environmental Change Research Centre, University College London, Pearson Building, Gower Street, London, WC1E 6BT, UK The spatial distribution of surface sediment diatoms was analyzed using ArcGIS in the Round Loch of Glenhead, an acid upland lake in south-west Scotland. The assemblages were composed almost entirely of benthic species. Tabellaria quadriseptata was fairly evenly distributed across the loch but some species (Navicula madumensis, Brachysira brebissonii, Aulacoseira perglabra and Eunotia vanheurckii var 1) showed rather patchy distributions. Ordination analysis was performed to assess the influence of environmental and spatial variables on the diatom composition of the samples. Loss of ignition was sig- nificantly negatively correlated with redundancy analysis species axis 1 (r = –0.77), indi- cating the influence of substrate on the diatom assemblages. The positive relationship be- tween theoretical bottom shear stress resulting from wind stress and redundancy analysis (r = 0.31) suggests wind stress also influences the spatial distribution of diatoms within the loch. Spatial variables [(principal coordinates of neighbour matrices (PCNM 1 and PCNM3) positively correlated with redundancy analysis axis 2], indicated that spatial variables, ignored in former studies, are a further influence on diatom distribution. Unique environmental and spatial variables explained 27.3% and 8.6% of diatom variability re- spectively. Environmental and spatial interactive variables combined explained 4.8% of variation. Although the pure contribution of spatial variables was only 8.6%, the study highlighted the importance of differences in the spatial distribution of different benthic di- atom species in this upland lake. Keywords: Diatom, distribution, composition, benthic, substrate, wind, lake, Scotland Abbreviations: DCA – detrended correspondence analysis, GAM – generalizes additive models, I – downward irradiance, Kd(PAR) – diffuse attenuation coefficient for down- ward irradiance, LOI – sediment organic matter content, PAR – photosyntheticall active radiation, RDA – redundancy analysis, TBSS – theoretical bottom shear stress ACTA BOT. CROAT. 68 (2), 2009 367 * Corresponding author: hongyanghy@gmail.com U:\ACTA BOTANICA\Acta-Botan 2-09\Yang.vp 6. listopad 2009 11:57:58 Color profile: Disabled Composite 150 lpi at 45 degrees Introduction It is often assumed that sediment deposition below a certain water depth is conformable and representative. However, the complexity of diatom deposition in lake basins is receiv- ing increasing attention as a result of multicore studies (ANDERSON 1989, 1990a, b; ANDER- SON 1998; ADLER and HUBENER 2007). In these studies, research focused on the selection of a representative core site from multiple cores collected from a range of water depths and lo- cations within a lake. Studies of transects across lakes from littoral to pelagic zones also demonstrated the variability in surface sediment diatom composition (MERILÄINEN 1971, BRADBURY and WINTER 1976, KINGSTON et al. 1983, KAUPPILA 1998). The majority of sites investigated spatially have been lowland lakes. In upland lakes, the spatial variability of living benthic diatom communities has been surveyed, but only from samples spatially limited to the littoral zone (JONES and FLOWER 1986) or, in the case of surface sediment diatom assemblages, from samples collected along transects (ALLOTT 1991). However, transect sampling is an imperfect method of cap- turing whole-lake distributions (WATTS and HALLIWELL 1996). A stratified random sample (SRS) method that ensures good area coverage and preserves the advantage of randomness is preferred (WATTS and HALLIWELL 1996). Diatom samples collected according to the SRS method can be analysed using multi- variate statistical techniques to identify the principal factors that explain the distribution. Until now, only environmental variables have been included in such analyses (KAUPPILA 1998). Here we also consider the importance of spatial variables. Spatial patterns in the abundance and distribution of organisms are inherent properties of ecological systems (FORTIN and DALE 2005), and therefore need to be included (LEGENDRE and FORTIN 1989, BORCARD et al. 1992). Consequently, we use principal coordinates of neighbour matrices (PCNM) a technique introduced by BORCARD and LEGENDRE (2002) and now used to con- struct spatial models in various fields of ecology (BORCARD et al. 2004, CRIST et al. 2006). The respective contributions of environmental and spatial variables to the variability of surface sediment diatom can then be calculated using the variation partitioning technique (BORCARD et al. 1992). Here we use this approach to analyse the spatial patterns of benthic diatoms in a Scottish upland lake, the Round Loch of Glenhead. Materials and methods Study site The Round Loch of Glenhead (British National Grid reference NX 450 804, 55°5’ N, 4°25’ W) is a small lake located in the Galloway hills of south-west Scotland (Fig. 1). The climate of Galloway is mild oceanic with an annual average precipitation of ca. 2500 mm yr–1 and an average temperature 8.5 °C with a range from –6 °C to 27 °C (YANG 2009). Soils in the catchment are mainly deep peats and peaty podsols; skeletal soils and bare rock occur on the steepest slopes (JONES et al. 1989). The lake itself is situated within a rugged moorland landscape at 295 m altitude and has an open water area of 12.5 ha with a mean water depth of 4.3 m (maximum depth is 14.0 m). The average pH of the lake (04/2005–03/2006) is 5.2 and dissolved organic carbon is 362.5 µmol L–1 (for more details of water chemistry, see SHILLAND et al. 2006). 368 ACTA BOT. CROAT. 68 (2), 2009 YANG H., FLOWER R. J., BATTARBEE R. W. U:\ACTA BOTANICA\Acta-Botan 2-09\Yang.vp 6. listopad 2009 11:57:58 Color profile: Disabled Composite 150 lpi at 45 degrees ACTA BOT. CROAT. 68 (2), 2009 369 SPATIAL DISTRIBUTION OF SURFACE SEDIMENT DIATOMS Fig. 1. Location and map of the Round Loch of Glenhead, south-west Scotland, UK, showing lake bathymetry, the main distribution areas of the dominant aquatic macrophytes (Isoetes lacustris, Juncus bulbosus var. fluitans, Lobelia dortmanna), loss on ignition (as histo- grams), the location of the weather station and the position of the ten stakes. Note that , which occurs sporadically in the loch, is not shown because its main growth period was after the main diatom sampling in April 2007. U:\ACTA BOTANICA\Acta-Botan 2-09\Yang.vp 6. listopad 2009 11:58:09 Color profile: Disabled Composite 150 lpi at 45 degrees Diatom habitats in the lake are confined to the littoral zone. They are dominated by the epilithon growing on stones and boulders that occur around the whole perimeter of the lake and the epiphyton growing principally on submerged plants. The dominants are Juncus bulbosus var. fluitans, Lobelia dortmanna, and Isoetes lacustris (SHILLAND et al. 2006). Juncus bulbosus var. fluitans is abundant in sandy bays, especially in the eastern littoral ar- eas, whereas Lobelia dortmanna occurs around almost all the margins of the lake and Isoetes lacustris grows on mud surfaces with water depths of up to 3 m (YANG 2009). The photic depth is about 6 m deep and epipsammic and epipelic diatoms can be found down to 6 m (YANG 2009). Sampling The stratified random sampling (SRS) method was used to collect the surface sediment samples in the loch from 22nd to 28th April 2007. The loch area was divided into 40 × 40 m grids (Fig. 1) and GPS positions (as OSGB 36) of the centre points of each grid were set up on a GARMIN GPS III. The boat was navigated according to the waypoints and anchored at the centre of each grid square. When the boat stopped drifting, one surface sample was collected randomly within each grid using either an epipelic sampler (YANG and FLOWER 2009), when the water depth was less than 1.5 m, or a Renberg corer (RENBERG 1991) when water depth was more than 2.0 m. The top 1 cm sediment was removed for surface sedi- ment samples. Shallow water samples (from ca. 0.5 m depth) were collected using the epipelic sampler at ten georeferenced points (Fig. 1) around the lake shore. All samples were transported back to the laboratory and stored at 4 °C. Diatom analysis Each sample was treated in the same way with no attempt to separate live and dead cells. Subsequently the samples were oxidised using hydrogen peroxide (H2O2) by heating in a water bath (RENBERG 1990). Microspheres were added to calculate the diatom concen- tration (BATTARBEE and KNEEN 1982). All samples were mounted on microscope slides us- ing Naphrax as a mountant. Diatom valves were identified and counted using a Leitz re- search microscope under oil immersion at 1000 times magnification and phase contrast. A minimum of 500 valves was identified and counted for each sample. The abundance contours of the dominant taxa were analysed using the IDW method in ArcGIS 9.0. Environmental data Environmental variables used in the analysis included photosynthetically active radia- tion (PAR), wind speed and direction, bottom shear stress and sediment organic matter con- tent (LOI). PAR was measured using a LICOR LI-250 underwater light meter at 0.5 m depth inter- vals between 10 am and 2 pm in the loch centre. Diffuse attenuation coefficients for down- ward irradiance Kd (PAR) were calculated from the slope of the linear regression of the nat- ural logarithm of downward irradiance (I) versus depth (Z), and values only from a fit r2 > 0.95 were accepted (KIRK 1994): 370 ACTA BOT. CROAT. 68 (2), 2009 YANG H., FLOWER R. J., BATTARBEE R. W. U:\ACTA BOTANICA\Acta-Botan 2-09\Yang.vp 6. listopad 2009 11:58:09 Color profile: Disabled Composite 150 lpi at 45 degrees Kd(PAR) = –1/z ln(Iz/I0) where Kd(PAR) is the diffuse attenuation coefficient for downward irradiance, Iz is the PAR measurement at depth z, and I0 is the PAR measurement at the water surface. The kd(PAR) and surface PAR (I0) were used to infer the PAR at different water depths. The temperatures at different water depths (1–12 m), wind speed and direction data were monitored by the weather station installed on a platform floating in the lake centre in October 2005. Data were provided by D. Monteith. The theoretical bottom shear stress (TBSS) t was calculated as the following formula (JAMES et al. 2004): t = H r u p( (2 / ) ) 2sinh(2 ) 3 0.5 T kh     where t is the calculated bottom shear stress, H is the wave height, r is the density of water, T is the wave period, u is the kinematic viscosity, k is the wave number (2p/L where L = wavelength, cm), and h is the water depth. Wave characteristics (H, T and L) were calcu- lated from wind speed, water depth and fetch distance using wave models according to the Coastal Engineering Research Center (1984). Wind speed was measured from weather sta- tion located near the loch centre and the one point measurement was cautiously extrapo- lated to the whole loch. For LOI, the percentage dry weights (DW) of sediment samples were calculated after drying a known weight of sample overnight at 105 °C and then combusting in a muffle fur- nace at 550 °C for 2 hours. Spatial data In the field, BNG (British National Grid) coordinates of each sampling site were re- corded. The original coordinate values were z-score transformed and these standard coor- dinates were used to create the dataset of spatial variables derived from PCNM using the program Spacemaker2 (Borcard and Legendre 2004, http://www.bio.umontreal.ca/legendre/). A matrix of Euclidean distances between samples was created and subsequently truncated based on truncation distance, which was equal to or larger than the largest distance between neighbours. Data analysis All environmental data were z-score transformed (LEGENDRE and GALLAGHER 2001) if the normality criteria were not satisfied. Species data were Hellinger transformed, because this method minimises the effects of the large number of zeros common in species abun- dance data (LEGENDRE and GALLAGHER 2001). Species with percent abundance of less than 3% per single sample were considered as »rare« taxa and were excluded from further ordi- nation analysis. Preliminary detrended correspondence analysis (DCA) was performed us- ing the program CANOCO 4.5 to measure the patterns of compositional variation and the biological species turnover (the gradient length) (ter BRAAK and PRENTICE 1988). Redun- dancy analysis or canonical correspondence analysis was selected based on gradient length to explore the relationships between diatom assemblages and environmental variables (TER ACTA BOT. CROAT. 68 (2), 2009 371 SPATIAL DISTRIBUTION OF SURFACE SEDIMENT DIATOMS U:\ACTA BOTANICA\Acta-Botan 2-09\Yang.vp 6. listopad 2009 12:10:03 Color profile: Disabled Composite 150 lpi at 45 degrees BRAAK and PRENTICE 1988). To reduce the possible effects of the difference in the number of variables included in each set of explanatory variables, we only used those environmen- tal variables that were significant based on a forward selection (p £ 0.05 after 999 random permutations) (ØKLAND and EILERTSEN 1994). To explore the relationship between single species distribution and environmental variables, GAM (HASTIE and TIBSHIRANI 1990) were performed using the R-language (R Development Core Team 2006) mgcv package (WOOD 2008). With respect to the spatial variables, the coordinates were created using PCNM and those eigenvectors with positive eigenvalues were retained for inclusion in the subsequent ordination analysis. Variation partitioning was used to quantify the proportion of the variation in diatom assemblage explained by variation in each of the combinations of environmental, and spatial variable sub-models (BORCARD et al. 1992). The unadjusted R2 value was biased (PERES-NETO et al. 2006) and therefore they were adjusted into R2a. Based on the proportion of variation explained (R2a) in the analyses, the contributions of every component to the total variation in the diatom assemblages were calculated (BORCARD et al. 1992). The significances of fractions were tested by means of 999 permutations under a re- duced model. All the analyses were performed using the R-language (R Development Core Team 2006) function varpart in the vegan package (OKSANEN et al. 2008). Results Distribution of surface sediment diatoms A total of 18 genera and 94 species were collected during the study. The mean diatom abundance was 5.802 × 103 valves DW mg–1 with a SD of 4.766 × 103 valves DW mg–1. Most species (80%) belonged to the genera Eunotia, Navicula, Frustulia or Tabellaria. The most abundant species were Navicula leptostriata Jørgensen, Frustulia rhomboides var. saxonica (Rabenhorst) de Toni, Eunotia incisa W. Sm. ex Greg., Tabellaria quadriseptata Knudson, Brachysira brebissonii R. Ross in Hartley, Aulacoseira perglabra Østrup and Eunotia vanheurckii var. 1 R.J.Flower. The environmental variables are summarised in ta- ble 1. GAM models that fitted the distribution of Navicula leptostriata abundance included LOI, PAR and TBSS and they were able to account for 54.8% of the variance. The selected models for Frustulia rhomboides var. saxonica and Eunotia incisa both included LOI and PAR, amounting to 35.9% and 24.9% of the total variance, respectively. Although the mod- els that fitted Tabellaria quadriseptata and Eunotia vanheurckii var 1 only included LOI, 372 ACTA BOT. CROAT. 68 (2), 2009 YANG H., FLOWER R. J., BATTARBEE R. W. Tab. 1. Summary Statistics for environmental variables. LOI=loss of ignition, PAR = photosynthe- tically active radiation, and TBSS= theoretical bottom shear stress. Min Max Mean SD Water depth 0.5 14.0 4.9 3.8 Dry weight 0.07 79.14 15.21 21.89 LOI 0.64 61.29 31.93 15.30 PAR 0.07 398.72 100.15 106.19 Temperature 9.7 11.3 10.7 0.5 TBSS 0 0.0221 0.0011 0.0041 U:\ACTA BOTANICA\Acta-Botan 2-09\Yang.vp 6. listopad 2009 11:58:09 Color profile: Disabled Composite 150 lpi at 45 degrees the models accounted for 40.7% and 32.6% of the total variance, respectively. The models that fitted Brachysira brebissonii and Aulacoseira perglabra included PAR and TBSS and they amounted to 17.8 and 24.0% of the total variance, respectively. The spatial distributions of the dominant taxa are shown in figures 2 and 3. Navicula leptostriata was common in the northern shallow water but was seldom found in the east- ern and southern basin (Fig. 2). Frustulia rhomboides var saxonica and Eunotia incisa oc- curred mostly in the middle and southern parts of the lake basin. Navicula madumensis (=Kobayasiella madumensis, LANGE-BERTALOT 1999) JØRGENSEN (1948) occurred mostly ACTA BOT. CROAT. 68 (2), 2009 373 SPATIAL DISTRIBUTION OF SURFACE SEDIMENT DIATOMS Fig. 2. Spatial distribution of surface sediment diatom concentrations (103 valves DW mg–1) in the Round Loch of Glenhead in April 2007. Upper left, Navicula leptostriata; upper right, Frustulia rhomboides var. saxonica; bottom left, Eunotia incisa; botttom right, Navicula madumensis. U:\ACTA BOTANICA\Acta-Botan 2-09\Yang.vp 6. listopad 2009 11:58:20 Color profile: Disabled Composite 150 lpi at 45 degrees in the eastern region of the loch. Tabellaria quadriseptata was widely distributed in the loch but with three small high abundance patches in the southern basin (Fig. 3). Brachysira brebissonii and Aulacoseira perglabra were concentrated in the loch centre. Eunotia vanheurckii var. 1 was mainly found in the northern and middle basin. Ordination analysis The gradient length of the first axis explored by DCA is 1.75 SD. This indicates that spe- cies turnover was at a range where linear species response models RDA could be suitable. 374 ACTA BOT. CROAT. 68 (2), 2009 YANG H., FLOWER R. J., BATTARBEE R. W. Fig. 3. Spatial distribution of surface sediment diatom concentrations (103 valves DW mg–1) in the Round Loch of Glenhead in April 2007. Upper left, Tabellaria quadriseptata; upper right, Brachysira brebissonii; bottom left, Aulacoseira perglabra; botttom right, Eunotia van- heurckii var. 1 U:\ACTA BOTANICA\Acta-Botan 2-09\Yang.vp 6. listopad 2009 11:58:36 Color profile: Disabled Composite 150 lpi at 45 degrees The environmental variables included dry weight, LOI, PAR, temperature and theoreti- cal bottom shear stress (TBSS). Forward selection identified LOI, PAR and TBSS as sig- nificant environmental variables for the later ordination analysis and variation partition methods. Spatial variables PCNM1, PCNM2, PCNM3 and PCNM4 were calculated and all four variables were included in the later analysis. RDA captures the variance in the spe- cies-environment and space relationship quite well; 77.3% by the first two axes (Tab. 2). The first ordination axis was negatively related to LOI (r = – 0.772) and PAR (r = –0.437). PAR, TBSS, PCNM1 and PCNM 3 were positively related to axis 2. Variation partitioning The variation in surface sediment diatom assemblages was partitioned into four parts: environmental variation (E), spatial variation (S), spatially structured environmental varia- tion (ES) and unexplained variation (U). Their respective percentages were 27.3%, 8.6%, 4.8% and 59.3%. Discussion Spatial distribution of surface sediment diatoms Georeferencing and organized sampling techniques are of major benefit in assessing the spatial distribution of modern diatoms in a lake (WATTS and HALLIWELL 1996). For the Round Loch of Glenhead these techniques have demonstrated major differences in the dis- ACTA BOT. CROAT. 68 (2), 2009 375 SPATIAL DISTRIBUTION OF SURFACE SEDIMENT DIATOMS Tab. 2. Summary statistics for the first four axes of the redundancy analysis (RDA) of diatom-envi- ronment and space with Hellinger transformation of diatom species, z – standardization of environmental variables and down-weighting of rare species. Only environmental variables selected in the forward selection procedure are presented in the ordination. TBSS = theoreti- cal bottom shear stress, LOI = loss of ignition, PAR= photosynthetically active radiation, PCNM = principal coordinates of neighbour matrices. Axes 1 2 3 4 Eigenvalue: 0.184 0.094 0.036 0.018 Species-environment and space correlations: 0.858 0.75 0.705 0.669 Cumulative percentage variance of species data: 18.4 27.8 31.3 33.1 of species-environment and space relation: 51.2 77.3 87.2 92.2 Weighted correlation with species axes TBSS 0.005 0.311 0.610 –0.050 PAR –0.437 0.456 –0.097 0.061 LOI –0.772 –0.203 –0.105 0.103 PCNM1 0.262 0.314 –0.390 0.406 PCNM2 0.196 0.019 0.312 0.327 PCNM3 –0.157 0.378 –0.181 –0.284 PCNM4 0.139 0.128 –0.188 –0.179 U:\ACTA BOTANICA\Acta-Botan 2-09\Yang.vp 6. listopad 2009 11:58:36 Color profile: Disabled Composite 150 lpi at 45 degrees tribution of benthic diatoms according to species and to spatial location. Tabellaria quadriseptata occurred almost evenly everywhere in the loch. Although it can be trans- ported around the lake basin, its wide spatial distribution indicated that it could be a habitat generalist taxon. On the other hand, there were clearly patchy distributions of Navicula madumensis, Brachysira brebissonii, Eunotia vanheurckii var. 1 and Aulacoseira pergla- bra. The reasons for the different spatial distribution of surface sediment diatom assem- blages are discussed below. In Lake Salkolanjärvi, results indicated that water depth was the main environmental variable controlling the distribution of surface sediment diatoms (KAUPPILA 1998). Argu- ably, water depth is an integrated variable, correlated directly or indirectly with light, tem- perature and other environmental variables. In this study, underwater light PAR and water temperature, rather than water depth, were considered. RDA indicated that there was a sig- nificantly negative correlation between LOI and RDA axis 1 (r = –0.77), suggesting the in- fluence of substrate on the diatom assemblages. Eunotia vanheurckii var. 1 is a typical epipsammic diatom in upland lochs (ALLOTT 1991). One of its aggregation areas was a sandy area in the loch (Fig. 3) and this patchy distribution probably supports the influence of substrate on the distribution of this diatom. PAR correlated to RDA axis 1 and 2. PAR is a major regulator of photosynthesis and consequently plays an important role in driving algal productivity and determining species composition in lakes (HILL 1996). TBSS correlated to axis 2, indicating the influence of wind stress on the spatial distribution of diatoms in the loch. Although the GAM that fitted the distribution of Tabellaria quadriseptata did not include TBSS, wind and wave action may be responsible for the relatively even distribution of this species. Aulacoseira per- glabra, a planktonic diatom species, is not found living in the loch today (e.g. JONES 1987, ALLOTT 1991, YANG 2009), but was abundant in the early Holocene (JONES 1987). The cur- rent occurrence of these diatoms in the deeper water surface sediments are probably the re- sult of sediment resuspension processes from the exposure of old sediment surfaces in this region of the loch. The GAM models that fitted Aulacoseira perglabra included TBSS, supporting this interpretation. Comparisons between the distribution of macrophytes and surface sediment diatoms show no clear relationship, probably because the epiphytic dia- toms were removed and re-distributed effectively by water turbulence. The influence of spatial variables Although the role of spatial variables is increasingly recognised in the ordination analy- sis of organism composition (e.g. WAGNER 2003, BORCARD et al. 2004, BEISNER et al. 2006), spatial variables have seldom been considered in the analysis of diatom distributions. Spa- tial structures observed in floristic composition can arise from: (1) autogenous structure (S), independent of any environmental variation; and (2) exogenous structure (ES), which results when species respond to environmental variables that are themselves spatially structured (FORTIN and DALE 2005, JONES et al. 2008). Before the introduction of the varia- tion partition technique (BORCARD et al. 1992), S and ES were not distinguished and both considered together as spatial variables in most studies. In the present study, we can extract the variation explained commonly by environmental and spatial variables (ES) by partial ordination analysis. The result indicates that 36% (=4.8/(4.8+8.6)) of the variation in the spatial sub-models was due to environmental interaction. The contribution of unique spa- 376 ACTA BOT. CROAT. 68 (2), 2009 YANG H., FLOWER R. J., BATTARBEE R. W. U:\ACTA BOTANICA\Acta-Botan 2-09\Yang.vp 6. listopad 2009 11:58:36 Color profile: Disabled Composite 150 lpi at 45 degrees tial variables to the total variability of the surface sediment diatom assemblage was only 8.6%. However, although this is a relatively small figure the result highlights the impor- tance of considering spatial variables separately from environmental variables in account- ing for variability in surface sediment diatom composition. The small value of total diatom variation explained by spatial structure alone (8.6%) suggests that no significant unique spatial-structuring of diatom variation was missed in the Round Loch of Glenhead. The positive relationship between both PCNM1 and PCNM3 and RDA species axis 2 (Tab. 2) confirmed the influence of spatial variables on surface sediment diatom composition. The unique spatial pattern (S) is more important to species that are habitat generalists or that ag- gregate in habitat patches (CRIST et al. 2006). Further research In this study, the combined contribution of environmental and spatial variables to the variation in diatom composition was only 40.7%. The high unexplained contribution may result from both insufficient measurement of variables and the limitations of using a single time period for sampling in a system that is quite dynamic (BEISNER et al. 2006). In addition the factors that control the distribution of dead cells from living ones will differ. In the case of dead cells it is important to understand the taphonomic processes that control the trans- port, deposition and resuspension of cells across the lake basin, and in the case of living cells diatom distributions may be controlled as much by seasonal variability in hydro- chemistry and by grazer suppression as by water depth and habitat (KOETSIER 2005). Iden- tifying the factors that control diatom distributions across lake basins with respect to both living and dead cells remains an important area for research (YANG et al, in preparation). This is relevant to attempts to improve the design of field sampling programmes, especially those that aim to characterise the floras of a whole lake rather than of a localised sampling site within a lake. Conclusions Study of the spatial distribution of surface sediment diatoms in the Round Loch of Glenhead has indicated that many taxa (Navicula madumensis, Brachysira brebissonii, Aulacoseira perglabra and Eunotia vanheurckii var. 1, in particular) were unevenly distrib- uted in the loch. The results of ordination analysis showed that LOI and PAR were nega- tively correlated to RDA diatom variation axis 1, while PAR, TBSS and the spatial vari- ables (PCNM 1 and PCNM3) were positively correlated with axis 2. These observations suggest that substrate, wind-induced water turbulence and the spatial configuration of the loch basin significantly influence benthic diatom distributions in the loch. We conclude that the distribution of diatom taxa is related to the habitats in the lake as well as to transport processes and sediment resuspension. In the case of Aulacoseira perglabra, resuspension of diatom valves from exposed old sediment surfaces is suspected. Unique spatial variables, not considered in former studies on diatom distribution, ex- plained 8.6% of the total variability of the surface sediment diatom composition. Although this value is not very high, the result confirmed the importance of spatial variables inde- pendent of environmental variables in explaining the variability in surface sediment diatom composition. ACTA BOT. CROAT. 68 (2), 2009 377 SPATIAL DISTRIBUTION OF SURFACE SEDIMENT DIATOMS U:\ACTA BOTANICA\Acta-Botan 2-09\Yang.vp 6. listopad 2009 11:58:36 Color profile: Disabled Composite 150 lpi at 45 degrees Acknowledgements We are grateful to Scottish Natural Heritage and the EU Euro-limpacs project (GOCE- -CT-2003-505540) for funds to support this study and to the Dorothy Hodgkin’s Postgradu- ate Award Scheme for supporting Hong Yang. Fieldwork was greatly assisted by support from the Forestry Commission staff especially by Neil Grieve, Sany White and Geoffrey Shaw, from the ECRC, UCL, especially James Shilland, Ewan Shilland, Simon Turner, Da- vid Hunt and Yan Zhao, from the Geography Department, UCL, especially Julian Thomp- son and Raja Singh, from Tim Allott, Manchester University, and all the others who helped with fieldwork. Laboratory work was greatly assisted by Janet Hope, Ian Patmore, and Tula Maxted. References ADLER, S., HUBENER, T., 2007: Spatial variability of diatom assemblages in surface lake sediments and its implications for transfer functions. Journal of Paleolimnology 37, 573–590. ALLOTT, T. E. H., 1991: The reversibility of lake acidification: a diatom study from the Round Loch of Glenhead, Galloway, Scotland, Department of Geography. University of London, London. ANDERSON, N. J., 1989: A whole-basin diatom accumulation rate for a small eutrophic lake in Northern Ireland and its palaeoecological implications. Journal of Ecology 77, 926–946. ANDERSON, N. J., 1990a: Spatial pattern of recent sediment and diatom accumulation in a small, monomictic, eutrophic lake. Journal of Paleolimnology 3, 143–160. ANDERSON, N. J., 1990b: Variability of sediment diatom assemblages in an upland, wind- -stressed lake (Loch Fleet, Galloway, S.W. Scotland). Journal of Paleolimnology 4, 43–59. ANDERSON, N. J., 1998: Variability of diatom-inferred phosphorus profiles in a small lake basin and its implications for histories of lake eutrophication. Journal of Paleo- limnology 20, 47–55. BATTARBEE, R. W., KNEEN, M. J., 1982: The use of electronically counted microspheres in absolute diatom analysis. Limnology and Oceanography 27, 184–188. BEISNER, B. E., PERES, P. R., LINDSTROM, E. S., BARNETT, A., LONGHI, M. L., 2006: The role of environmental and spatial processes in structuring lake communities from bacteria to fish. Ecology 87, 2985–2991. BORCARD, D., LEGENDRE, P., 2002: All-scale spatial analysis of ecological data by means of principal coordinates of neighbour matrices. Ecological Modelling 153, 51–68. BORCARD, D., LEGENDRE, P., 2004: SpaceMaker2 – User’s guide. Département de sciences biologiques, Université de Montréal. BORCARD, D., LEGENDRE, P., AVOIS-JACQUET, C., TUOMISTO, H., 2004: Dissecting the spa- tial structure of ecological data at multiple scales. Ecology 85, 1826–1832. BORCARD, D., LEGENDRE, P., DRAPEAU, P., 1992: Partialling out the spatial component of ecological variation. Ecology 73, 1045–1055. 378 ACTA BOT. CROAT. 68 (2), 2009 YANG H., FLOWER R. J., BATTARBEE R. W. U:\ACTA BOTANICA\Acta-Botan 2-09\Yang.vp 6. listopad 2009 11:58:36 Color profile: Disabled Composite 150 lpi at 45 degrees BRADBURY, J. P., WINTER, T. C., 1976: Areal distribution and stratigraphy of diatoms in the sediments of Lake Sallie, Minnesota. Ecology 57, 1005–1014. COASTAL ENGINEERING RESEARCH CENTER, 1984: Shore protection manual, Volume 1, U. S. Army Corps of Engineers, Waterways Experiment Station, Vicksburg, Mississippi, USA. CRIST, T. O., PRADHAN-DEVARE, S. V., SUMMERVILLE, K. S., 2006: Spatial variation in insect community and species responses to habitat loss and plant community composition. Oecologia 147, 510–521. FORTIN, M. J., DALE, M. R. T., 2005: Spatial analysis: a guide for ecologists. Cambridge University Press, Cambridge. HASTIE, T. J., TIBSHIRANI, R. J., 1990: Generalized additive models. Chapman and Hall, London. HILL, W. R., 1996: Effects of light. In: STEVENSON, R. J., BOTHWELL, M. L., LOWE, R. L. (ed.), Algal ecology: freshwater benthic ecosystems, 121–148. Academic Press, San Diego. JAMES, W. F., BARKO, J. W., BUTLER, M. G., 2004: Shear stress and sediment resuspension in relation to submersed macrophyte biomass. Hydrobiologia 515, 181–191. JONES, M. M., TUOMISTO, H., BORCARD, D., LEGENDRE, P., CLARK, D. B., OLIVAS, P. C., 2008: Explaining variation in tropical plant community composition: influence of envi- ronmental and spatial data quality. Oecologia 155, 593–604. JONES, V. J., 1987: A palaeoecological study of the post-glacial acidification of the Round Loch of Glenhead and its catchment. Department of Geography. University of London, London. JONES, V. J., FLOWER, R. J., 1986: Spatial and temporal variability in periphytic diatom com- munities: Palaeoecological significance in an acidified lake. In: SMOL, J. P., BATTARBEE, R. W., DAVIS, R. B., MERILÄINEN, J. (ed.), Diatoms and lake acidity: reconstructing pH from siliceous algal remains in lake sediments, 87–94. Junk Publishers, Dordrecht. JONES, V. J., STEVENSON, A. C., BATTARBEE, R. W., 1989: Acidification of lakes in Galloway, south west Scotland – a diatom and pollen study of the post-glacial history of the Round Loch of Glenhead. Journal of Ecology 77, 1–23. KAUPPILA, T., 1998: Variability of surface sediment diatom assemblages in Lake Salko- lanjaervi, Finnland. Proceedings 15 International diatom symposium, Perth, 263–274. KINGSTON, J. C., LOWE, R. L., STOERMER, E. F., LADEWSKI, T. B., 1983: Spatial and temporal distribution of benthic diatoms in Northern Lake-Michigan. Ecology 64, 1566–1580. KIRK, J. T., 1994: Light and photosynthesis in aquatic ecosystems. Cambridge University Press, New York. KOETSIER, P., 2005: Response of a stream diatom community to top predator manipula- tions. Aquatic Sciences 67, 517–527. LEGENDRE, P., FORTIN, M. J., 1989: Spatial pattern and ecological analysis. Plant Ecology 80, 107–138. LEGENDRE, P., GALLAGHER, E. D., 2001: Ecologically meaningful transformations for ordi- nation of species data. Oecologia 129, 271–280. ACTA BOT. CROAT. 68 (2), 2009 379 SPATIAL DISTRIBUTION OF SURFACE SEDIMENT DIATOMS U:\ACTA BOTANICA\Acta-Botan 2-09\Yang.vp 6. listopad 2009 11:58:36 Color profile: Disabled Composite 150 lpi at 45 degrees MERILÄINEN, J., 1971: The recent sedimentation of diatom frustules in four meromictic lakes. Annales Botanici Fennici 8, 160–176. ØKLAND, R. H., EILERTSEN, O., 1994: Canonical correspondence-analysis with variation partitioning – some comments and an application. Journal of Vegetation Science 5, 117–126. OKSANEN, J., KINDT, R., LEGENDRE, P., O’HARA, B., SIMPSON, G. L., STEVENS, M. H. H., 2008. Vegan: Community Ecology Package 1.11–4 (http://cran.r-project.org/, http:// vegan.r-forge.r-project.org/). PERES-NETO, P. R., LEGENDRE, P., DRAY, S., BORCARD, D., 2006: Variation partitioning of species data matrices: Estimation and comparison of fractions. Ecology 87, 2614– 2625. R DEVELOPMENT CORE TEAM, 2006. R: A language and environment for statistical comput- ing. R Foundation for Statistical Computing, Vienna, http://www.R-project.org. RENBERG, I., 1991: The HON-Kajak sediment corer. Journal of Paleolimnology 6, 167– 170. RENBERG, I., 1990: A procedure for preparing large sets of diatom slides from sediment cores. Journal of Paleolimnology 4, 87–90. SHILLAND, E. M., MONTEITH, D. T., BONJEAN, M., BEAUMONT, W. R. C., 2006: The United Kingdom acid waters monitoring network data report for 2005–2006. Report to the De- partment of the Environment, Food and Rural Affairs. ENSIS Publishing, London. TER BRAAK, C. J. F., PRENTICE, I. C., 1988: A theory of gradient analysis. Advances in Eco- logical Research 18, 271–317. WAGNER, H. H., 2003: Spatial covariance in plant communities: Integrating ordination, geostatistics, and variance testing. Ecology 84, 1045–1057. WATTS, S., HALLIWELL, L., 1996: Essential environmental science: Methods and tech- niques. Routledge, London. WOOD, S. N., 2008. mgcv: GAMs with GCV smoothness estimation and GAMMs by REML/PQL, Available at http://www.R-project.org, http://cran.r-project.org/ web/ packages/mgcv/index.html. YANG, H., 2009. Spatial and temporal patterns of benthic diatoms in an upland Scottish loch (the Round Loch of Glenhead): with reference to recent environmental changes. PhD thesis, Department of Geography. University College London, London. YANG, H., FLOWER, R ., J., 2009: A portable hand-operated sampler for shallow water sur- face sediments with special reference to epipelic communities. Journal of Paleolim- nology 42, 317–324. 380 ACTA BOT. CROAT. 68 (2), 2009 YANG H., FLOWER R. J., BATTARBEE R. W. U:\ACTA BOTANICA\Acta-Botan 2-09\Yang.vp 12. listopad 2009 11:43:06 Color profile: Disabled Composite 150 lpi at 45 degrees