Richard Telford | University of Bergen (original) (raw)

Uploads

Papers by Richard Telford

Quaternary Science Reviews

Background / Purpose: Individual components of fragmentation – decreased patch size, increased di... more Background / Purpose: Individual components of fragmentation – decreased patch size, increased distance from source populations, increased patch disturbance and impermeability of intervening matrix – rarely occur in isolation. We compared bird community composition in forest fragments around the Budongo Forest Reserve in western Uganda. Our prediction was that some guilds would be more affected by certain aspects of fragmentation than others. Main conclusion: We found that different fragmentation impacts have something of a cumulative effect, especially on certain guilds. Species richness and diversity measures do not show fragmentation effects.

Climate of the Past, 2014

The so-called "8.2 ka event" is widely regarded as a major Holocene climate perturbation. It is m... more The so-called "8.2 ka event" is widely regarded as a major Holocene climate perturbation. It is most readily identifiable in the oxygen-isotope records from Greenland ice cores as an approximately 160-year-long cold interval between 8250 and 8090 years BP. The prevailing view has been that the cooling over Greenland, and potentially over the northern North Atlantic at least, was triggered by the catastrophic final drainage of the Agassiz-Ojibway proglacial lake as part of the remnant Laurentide Ice Sheet collapsed over Hudson Bay at around 8420 ± 80 years BP. The consequent freshening of surface waters in the northern North Atlantic Ocean and the Nordic Seas resulted in weaker overturning, and hence reduced northward ocean heat transport. We have reconstructed variations in the strength of the east-

IOP Conference Series: Earth and Environmental Science, 2009

This article was submitted without an abstract, please refer to the full-text PDF file.

Review of Palaeobotany and Palynology, 2014

A modern pollen dataset from China and Mongolia (18-52°N, 74-132°E) is investigated for its poten... more A modern pollen dataset from China and Mongolia (18-52°N, 74-132°E) is investigated for its potential use in climate reconstructions. The dataset includes 2559 samples, 229 terrestrial pollen taxa and four climatic variablesmean annual precipitation (P ann ): 35-2091 mm, mean annual temperature (T ann ): -12.1-25.8°C, mean temperature in the coldest month (Mt co ): -33.8-21.7°C, and mean temperature in the warmest month (Mt wa ): 0.3-29.8°C. Modern pollen-climate relationships are assessed using canonical correspondence analysis (CCA), Huisman-Olff-Fresco (HOF) models, the modern analogue technique (MAT), and weighted averaging partial least squares (WA-PLS). Results indicate that P ann is the most important climatic determinant of pollen distribution and the most promising climate variable for reconstructions, as assessed by the coefficient of determination between observed and predicted environmental values (r 2 ) and root mean square error of prediction (RMSEP). Mt co and Mt wa may be reconstructed too, but with caution. Samples from different depositional environments influence the performance of cross-validation differently, with samples from lake sediment-surfaces and moss polsters having the best fit with the lowest RMSEP. The better model performances of MAT are most probably caused by spatial autocorrelation. Accordingly, the WA-PLS models of this dataset are deemed most suitable for reconstructing past climate quantitatively because of their more reliable predictive power.

The Holocene, 2014

Transfer functions are widely used in palaeoecology to provide quantitative environmental reconst... more Transfer functions are widely used in palaeoecology to provide quantitative environmental reconstructions using biological proxies. Most models use all but the rarest taxa present in the training set, even though many may be unrelated to the environmental variable of interest. We hypothesise that retaining such non-informative taxa will reduce model robustness and present a method for variable selection motivated by the statistical learning algorithm in random forests. We apply our species-pruning algorithm into weighted averaging (WA) and maximum likelihood calibration of response curves (MLRCs), and compare results of boosted regression trees (BRTs) using artificial and real datasets. Results from the artificial data show that WA is particularly sensitive to the influence of both non-informative taxa and secondary environmental variables in the training set or fossil assemblage, and that BRTs are relatively immune to these effects. Furthermore, species-pruned WA and MLRCs offer substantial improvements over all-species models when the training set includes non-informative taxa but does not guard against confounding effects when species have bi-or multivariate responses to the primary and one or more secondary variables. Tests with a limited set of examples of real data indicate that BRTs, MLRCs or species-pruned models have no apparent advantage over WA. We discuss possible reasons for this contradiction and suggest that more tests are needed to properly evaluate BRTs and species-pruned models.

Biotropica, 2014

Seed banks contribute to forest regeneration after disturbance, but less is known about fern spor... more Seed banks contribute to forest regeneration after disturbance, but less is known about fern spore banks, particularly in a paleotropical context. We sampled the buried seed and fern spore bank in Mabira Forest, a 300 km 2 forest in central Uganda, to explore the effect of time since disturbance. Soil cores (5 cm depth) were taken from 39 plots across three different classes of 'recovery': (1) not disturbed since 1950; (2) logged between 1950 and 1980; and (3) cleared for agriculture between 1970 and 1990 but reforested since. Plant emergence was monitored in a glasshouse. We predicted that the seed bank would reflect time since disturbance, with more pioneer species in recently disturbed stands, and that the fern spore bank would reflect stand age less closely due to greater dispersal capacity. We recorded a median 752 seeds per square meter, most of which were trees; the most abundant species was the invasive tree Broussonetia papyrifera. The fern spore bank was twice as dense, but 95 percent of fern spores were of one species, Christella parasitica. Tree seed density was significantly affected by time since disturbance with fewer seeds in the older stands. Herb seed density, fern spore density, and species richness for all groups were not significantly affected by time since disturbance. Neither seed bank nor fern spore bank closely resembled the aboveground vegetation. We compared our results to existing literature on seed banks in tropical forests, finding that our densities are relatively high for African forests, but low compared to the Neotropics and Australia.

Series in Machine Perception and Artificial Intelligence, 2002

ABSTRACT A large fraction of the estimates of LGM sea-surface temperatures are based on planktoni... more ABSTRACT A large fraction of the estimates of LGM sea-surface temperatures are based on planktonic foraminifera assemblages and calculated using transfer functions. Despite the wide depth distribution of planktonic foraminifera in the upper ocean, transfer functions for foraminifera assemblages are usually calibrated against 10m SST. Recent work has shown that calibrating foraminifera assemblages against different depths can yield markedly different reconstructions, and that 10m reconstructions are rarely the most plausible as they typically don't explain the most variance in a time series of fossil data. These problems are most severe in the tropics. The tropics also have many LGM assemblages without good modern analogues, and the thermal profile of tropical CMIP5 LGM grid boxes lack analogues in the CMIP5 PI ocean. In view of these issues, we recalibrate the Atlantic planktonic foraminifera training sets against different depths, and reconstruct LGM temperatures. We show that the direction of east-west gradient in tropical LGM temperature anomalies is dependent on the calibration depth. If the most applicable calibration depths are deeper in the western tropical Atlantic than the eastern, as suggested by analysis of time series, we can reconstruct uniform cooling in the tropical Atlantic, greater than that previously reconstructed. We compare our new reconstructions with CMIP5 LGM temperatures.

We present a method to test the statistical significance of a quantitative palaeoenvironmental re... more We present a method to test the statistical significance of a quantitative palaeoenvironmental reconstruction inferred from biotic assemblages with a transfer function. A reconstruction is considered statistically significant if it explains more of the variance in the fossil data than most reconstructions derived from transfer functions trained on random environmental data. Given reconstructions of several environmental variables from the same fossil proxy, the method can determine which is the best reconstruction, and if there is sufficient information in the proxy data to support multiple independent reconstructions. Reconstructions that fail this test have limited credibility and should be interpreted with considerable caution.