Open AccessQuantitative forest‐composition sensing characteristics of pollen samples from Swedish lakes
Author(s) -
PRENTICE I. COLIN,
BERGLUND BJÖRN E.,
OLSSON TOMMY
Publication year - 1987
Publication title -
boreas
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.95
H-Index - 74
eISSN - 1502-3885
pISSN - 0300-9483
DOI - 10.1111/j.1502-3885.1987.tb00753.x
Subject(s) - pollen , picea abies , physical geography , taxon , sampling (signal processing) , forestry , geology , fagus sylvatica , environmental science , ecology , geography , biology , beech , filter (signal processing) , computer science , computer vision
Prentice, I. Colin, Berglund, Björn E. & Olsson, Tommy 1987 03 01: Quantitative forest‐composition sensing characteristics of pollen samples from Swedish lakes. Boreas , Vol. 16, pp. 43–54. Oslo. ISSN 0300–9483. Surface pollen percentages of major tree genera in 53 moderate‐sized lakes in south and central Sweden were compared with percentages of mean forest volume derived from survey plots within 5, 10, 20, 50 and 100 km of each lake. A maximum likelihood extended R‐value method was used to estimate relative slopes and intercepts of the pollen‐tree relationship at each forest sampling radius. Slopes generally went up and intercepts down as sampling radius was increased. Visual goodness‐of‐fit was optimal at 5–10 km for Picea and at 50–100 km for Pinus , consistent with theoretical pollen source areas. Fagus gave good visual fits at all radii, but Quercus gave indifferent fits and Betula and Alnus poor fits. The taxa differed in relative pollen representation, with relative R‐values on the order of 0.11 for Picea , 0.90 for Fagus , 1.0 for Pinus , 1.5 for Quercus , 2.3 for Alnus and 2.5 for Betula . Regional forest patterns can be reconstructed from lake pollen spectra by applying R‐values, but the ‘region’ represented depends on each taxon's pollen dispersal characteristics and spatial pattern. The R‐values given here could be used to calibrate European Holocene pollen spectra from moderate‐sized lakes for long‐term forest dynamic mapping.