Testing the palaeoclimatic signal from peat bogs - temperature or precipitation forcing?

The aim of this project was to compare a peatland proxy climate record with a temperature reconstruction based on chironomids in the sediments of a nearby lake in an attempt to assess the degree to which the peat-based record reflects temperature changes. Peat-based palaeoclimatic proxies measure bog surface wetness, which depends upon effective precipitation (precipitation minus evapotranspiration), and previous work involving comparisons with Little Ice Age records, and the analysis of long documentary climate records, had suggested the hypothesis that the temperature signal is more coherent and is dominant over the spatially and temporally incoherent precipitation signal. A 3 metre peat core of from Walton Moss, Cumbria, was analysed for plant macrofossils, colorimetric humification, and testate amoebae, to yield three independent climate proxies. A 240 cm core of sediments from Talkin Tarn, nine km south of Walton Moss, was analysed for chironomid head capsules, and air temperatures were reconstructed using a mean July temperature transfer function. The cores were dated using AMS radiocarbon dates and an Icelandic tephra. The three proxy reconstructions from Walton Moss show coherent patterns of change over the last 3000 years. The plant macrofossil data show the profile to be sensitive to mire water table fluctuations, with the numerous occurrences of the hygrophilous Sphagnum section Cuspidata modelled as low DCA scores, indicating climatic deteriorations. The main climatic deteriorations identified from these proxy data occur between cal. 2900-2830 BP; 2630-2590 BP; 1730 BP; 1550-1400 BP; 1120 BP; 840 BP; 680BP; 500-400 BP, and 240-150 BP. The chironomids from the sediments of Talkin Tarn showed changing domination of various taxa over time, with over 50 taxa identified. The fossil data were modelled with a transfer function from 153 Norwegian lakes (S. Brooks & H. J. B. Birks, pers. comm.), producing a maximum range of reconstructed values between 14.8-12.6°C, with an RMSEP of 1.0592. The mud/water interface was not retained by the core, although the reconstructed temperature for the sample at 10 cm (closest to the modern day) is 14.8°C, an exact match with the present recorded 30 year average mean July temperature of 14.8°C for Talkin Tarn. The results indicate that the midges appear to have responded to fluctuations in past temperature, most notably around cal. 2500 BP, when there is a significant increase in the abundance of Tanytarsus lugens group indicating a relatively large decrease in temperature of over 2°C. Other notable fluctuations occur earlier in the Holocene, illustrated by the decline in temperature from 14.8 to 13.6°C between ca. cal. 5300 and 4550 BP, and the lowering of temperature by ca. 1.3°C between 2000 and 1500 BP. Although the general trend of temperature fluctuations can be reconstructed from the chironomid transfer function, the comparatively low sampling resolution of the chironomids (one sample every ca. 250 years) prevents detailed inferences being made with respect to the rates and magnitudes of change over time. However, it is significant that the age/depth model for Talkin Tarn dates the main temperature decline to around cal. 2600 BP. A well known climatic deterioration has been recorded from many bogs around this time, as well as from a chironomid reconstruction in the Cairngorms. The Walton Moss data also show a climatic deterioration at this time. The other major temperature decline in the chironomid-inferred temperature reconstruction matches a general trend within the mire water table reconstructions from the testate amoebae data. A 1.3°C drop in temperature at Talkin Tarn is associated with an initial 7.4 cm rise in water table depth at Walton Moss at cal. 1730 BP, and comparatively high water tables continue (with subtle fluctuations, as indicated by the other proxy reconstructions) until cal. 920 BP, when they fall to below 11 cm. This corresponds with a 1°C increase in temperature at Talkin Tarn between cal. 1090-830 BP, and could be an indication of climatic change associated with the Medieval Warm Period. Although the three proxy reconstructions from Walton Moss detect changes associated with the Little Ice Age, there is no such indication from the chironomid-inferred temperatures. This may be due to the low sampling resolution, or may reflect other changes in the lake catchment, such as anthropogenic activity. Human influence is indicated by an increase in pastoral pollen indicators towards the top of the fossil pollen spectra (S. Morriss, pers. comm). It may also be suggested that chironomids respond not only to the relatively small changes of temperature within the Holocene, but also to changes in the pH of the lake water. However, this should only be a possible factor in lakes with acid rock catchments (S. Brooks, pers.comm.) which is not the case at Talkin Tarn. This project was designed as a palaeoecological test of the hypothesis that changes in bog surface wetness are driven primarily by temperature, and this hypothesis has been upheld. The project has further demonstrated that chironomids can successfully detect small scale temperature changes during the Holocene, based upon the new 153-lake transfer function, which also successfully models the present July temperature. The use of three independent proxies on the same peat profile gives a firmer foundation for proxy climate reconstructions – although time-consuming it is to be recommended for critical applications.