Palaeolimnological Investigation of Scottish Freshwater Lochs

March 2001


  1. This is the final report to SNIFFER on: Palaeolimnological Investigations of Scottish Freshwater Lochs. The project employs palaeolimnological techniques to evaluate the extent of eutrophication and/or acidification at 29 selected Scottish freshwater lochs of high environmental value or showing some signs of degradation. Two different basins were studied in two of the lochs, resulting in a total of 31 study sites.
  2. This report describes the lithostratigraphy, radiometric dating, and fossil diatom assemblages in a sediment core from each loch. Two levels of diatom analysis were employed: a relatively high resolution (15-20 samples) at those lochs considered to be of high interest, and a lower resolution (4-5 samples) at the remaining sites.
  3. Transfer functions were applied to the diatom data to generate reconstructions of total phosphorus (TP) concentrations and/or pH, following taxonomic harmonisation between the training sets and core species data. The TP reconstructions were calculated using either a Northwest European training set of 152 lakes (Bennion et al., 1996a) or an unpublished training set of 56 larger, less productive lakes, as appropriate. The pH reconstructions were produced using the Surface Water Acidification Programme (SWAP) training set of 167 lakes from largely acid-sensitive, upland regions of the UK, Norway and Sweden (Stevenson et al., 1991).
  4. A comparison of baseline diatom-inferred TP (DI-TP) concentrations with TP concentrations hindcasted from the PLUS export coefficient model (Ferrier et al., 1996, 1997b) for around 1850 AD was made at a number of the lochs. The difference between the baseline values calculated by the palaeolimnological and export-coefficient methods was assessed.
  5. The ratio of current measured TP to baseline TP calculated by the PLUS and diatom models was used to classify the lochs according to the Scottish standing waters classification scheme (Fozzard et al., 1999). Where there was a large discrepancy between current measured TP and DI-TP current, the ratio of DI-TP current to DI-TP baseline was also calculated to derive more meaningful output. The effect on the class banding and outputs from the classification scheme were examined, and the reasons for any discrepancies between the results of the two models were explored.
  6. The diatom-inferred pH (DI-pH) values calculated for the acid sensitive lochs could not be directly compared with the acidification criteria used in the Scottish standing waters classification scheme (Fozzard et al., 1999) as the scheme is based on changes in acid neutralising capacity (ANC) rather than loch pH. The DI-pH results did, however, provide additional information on lake acidification history and its impact on the lake biota, and were used to support the classification bandings. For Lochs Chon, Dee, Doon and Grannoch, the DI-pH values were compared with pH values hindcasted by the MAGIC model (Jenkins et al., 1997).
  7. A number of problems encountered in this project are summarised in the report, and include for some lochs: discrepancies between current measured TP and DI-TP current, lack of sensitivity of the diatom TP transfer function to reconstruct changes in situations where the species shifts are very subtle or where non-planktonic Fragilaria taxa dominate the assemblages, radiometric dating difficulties especially in the shallow lochs, insufficiently long sediment records to extend back to the desired baseline date of c. 1850 AD, and poor diatom preservation.
  8. Of the 31 study sites, the palaeolimnological investigations indicate that 19 (61%) have experienced eutrophication or some degree of enrichment, 7 (23%) have experienced acidification, one site (3%) exhibited signs of early acidification followed by recent enrichment, 2 (6.5%) sites have experienced no change, and for 2 (6.5%) sites the results were inconclusive owing to diatom preservation problems.
  9. The diatom-TP transfer functions were applied to 26 sites and resulted in a TP banding of class 1 for 17 (65%) of these, and class 2 for the remaining 9 (35%) lochs. While many sites demonstrated some degree of enrichment, only nine of these were downgraded to class 2 on the basis of DI-TP increases. The remaining lochs which exhibited signs of enrichment experienced changes of less than two-fold and, therefore, they were classified as class 1.
  10. The DI-pH transfer function was applied to 8 acid-sensitive sites and showed a pH decline of <0.5 pH units at 3 sites, a decline of 0.5-1.0 pH units at 3 sites and a decline of 1.0-1.5 pH units at 2 sites. The data show that all 8 lochs susceptible to acidification have experienced a gradual pH decline since around 1850 with a sharper decline since the mid-1900s in some cases associated with the exacerbating effect of afforestation. In contrast, the Scottish classification scheme based on ANC change highlights only Lochs Chon, Doon, Shiel and Grannoch as sites where acidification has led to a downgrading to either class 2 and 3, whilst Lochs Dee, Eck, Einich and Muick fall into class 1.
  11. A comparison of the baseline TP concentrations derived using the diatom and PLUS models showed a good match at some lochs while at others the results differed substantially. There was no systematic bias in the results. The greatest discrepancies occurred at the more productive lochs where the diatom model produced substantially higher baseline values than the PLUS model. The only lochs where the diatom model produced substantially lower baseline TP values than the PLUS model were shallow, macrophyte-dominated systems. For the large, deep, less productive lochs the models were in extremely good agreement.
  12. Of the 19 sites for which the ratios between current measured TP and baseline TP were calculated, both models produced TP class 1 at 7 (37%) sites and TP class 2 at 4 (21%) sites, the PLUS model produced TP class 1 and the diatom model TP class 2 at 3 (16%) sites, and vice versa at 5 (26%) sites. Given all possible error sources and that the models are based on entirely different methodologies there was remarkably good concordance between the TP values generated from the two approaches.
  13. Detrended correspondance analysis (DCA) proved to be a useful technique for illustrating patterns in the diatom data and for deriving ecological targets. Similarity in the pre-disturbance diatom assemblages of the large, deep lochs indicates that they were very similar ecologically and chemically in their unimpacted state. Species shifts to a new nutrient-rich flora in response to eutrophication and a new acid flora in response to acidification were observed in many of the lochs. One exception to this was Loch Grannoch whose pre-acidification flora lacked the circumneutral assemblage seen in the other deep lochs, and instead was comprised of acid-tolerant taxa, switching to a strongly acid flora as acidification progressed. This indicates that Loch Grannoch is naturally a more acid system and has never supported a circumneutral diatom flora.
  14. The diatom assemblages of many of the shallow lochs in the study were also very similar to each other but differed markedly from those in the deep lochs. The nature of the diatom assemblages reflects not only the different morphometry but also the greater range of habitats and the higher productivity and alkalinity of the shallow, lowland waters. The unimpacted state of the shallow lochs differs from the unimpacted state of the deep lochs both in terms of water chemistry and biota. The pre-disturbance assemblages and the species shifts, however, varied to a greater extent between the shallow lochs than that observed in the deep systems, reflecting the greater complexity of shallow systems.
  15. The study shows that many of the lochs experienced long periods with relatively stable diatom assemblages followed by recent periods of greater variability. For the acidified lochs, circa 1850 appears to be the time at which pH began to steadily decline and thus 1850 provides a sensible baseline against which to assess change. In many of the enriched lochs, the major species shifts have occurred in the last two to three decades resulting from increased nutrient inputs which vary on a site by site basis but include diffuse agricultural sources, sewage treatment works, and fish farms. Thus whilst 1850 acts as a baseline date against which to monitor change, it could be argued that a date of around 1960/1970 provides a more realistic target for restoration. It is highly unlikely that the nutrient concentrations and lake biota can be restored to conditions that existed at around 1850 when catchment land-use and management practices were very different from those in place today.
  16. The study illustrates the importance of palaeolimnological investigations for assessing change in the large, deep, currently oligotrophic waters and for providing an early warning signal. The species shifts towards the top of the Loch Lomond and Loch Awe cores clearly show that these lochs are starting to show signs of ecological change indicative of enrichment. These findings should be incorporated into management strategies for these lochs to prevent movement from class 1. Palaeolimnology may, therefore, be particularly valuable for detecting changes in lakes where water quality problems are not yet evident to allow management strategies to focus on protection and prevention rather than restoration.
  17. The presence of dense, inorganic sediment layers often coincident with short episodes of enhanced sediment accumulation rates was seen in some cores, e.g. Lochs Awe, Butterstone, Earn and Lomond. These layers are most likely to be derived from some kind of disturbance in the catchment such as ploughing, road construction or erosion resulting from forestry activities. However, the physical disturbance appeared to have no major impact on water quality as the diatom species shifts either substantially pre-dated or post-dated the minerogenic horizons.
  18. The diatom-TP results suggest that the Scottish classification scheme may be underestimating the true extent of impact on water quality and ecological change by setting the class 1 upper limit for TP change too high. Under the current system, lochs whose water quality and ecology are unaffected by human activity fall into the same class (class 1) as those lochs which are starting to show signs of ecological change and in some cases lochs which exhibit more marked changes indicative of enrichment. This may result in a number of sites which are showing signs of change from being overlooked. Likewise, the diatom-pH results suggest that the ANC bandings may also be set too high to detect those sites where pH has declined quite markedly. This suggests that perhaps the sensitivity of the scheme should be increased. Whilst it could be argued that all of the class 1 lochs have not been “significantly altered”, it may nevertheless be important for management purposes to distinguish between unimpacted and slightly impacted waters. Importantly, the scheme does appear to be adequate for identifying those lochs which have been more seriously impacted by either eutrophication or acidification.
  19. It is recommended that for high priority lochs, the best management tool is a combined approach: the export coefficient model for establishing the origins and changes in nutrient loadings, and the diatom record to reflect the nature and timing of the ecological response. The role of these techniques in defining reference conditions as required by the recently agreed EU Water Framework Directive is discussed.
  20. The final chapter makes a number of recommendations for further work and monitoring of Scottish freshwater lochs.

KEY WORDS: Palaeolimnology, Scottish lochs, eutrophication, acidification, diatoms, transfer functions, hindcasting, baselines, ecological targets, trajectories, water quality classification, lake management and conservation, Water Framework Directive.

Copies of this report are available from the Foundation price 50.00, less 20% to FWR members.