A Bayesian multinomial Gaussian response model for organism-based environmental reconstruction

• Published in: Journal of paleolimnology Vol. 24; no. 3; pp. 243 - 250
• Main Authors: Vasko, Kari, Toivonen, Hannu Tt, Korhola, Atte
• Format: Journal Article
• Language: English
• Published: Dordrecht Springer Nature B.V 01.09.2000
• Subjects: Calibration; Correlation coefficient; Environmental modeling; Freshwater; Studies
• ISSN: 0921-2728; 1573-0417
• DOI: 10.1023/A:1008180500301

We present a Bayesian hierarchical multinomial regression model (Bummer) for organism-based quantitative paleoenvironmental reconstruction. The model is based on the classical (direct) approach to calibration and on careful statistical environmental modeling that takes account of statistical dependencies among species. We compare our Bayesian model Bummer to seven other methods, including the widely used weighted averaging (WA) techniques and our previous Bayesian model Bum. The methods are evaluated on a surface-sediment chironomid training set of 62 subarctic lakes in northern Fennoscandia by comparing the cross-validation prediction statistics of different models. Bummer outperformed other methods, yielding the smallest prediction error, the smallest bias, and the largest correlation coefficient. We conclude that the promising performance of our Bayesian multinomial Gaussian response model is due to the following reasons: (i) the uncertainty concerning site specific latent variables is taken into consideration; (ii) ecological background knowledge is embedded to the model description; (iii) the species compositions are considered as a whole; and (iv) reconstruction is based on the classical approach to calibration.[PUBLICATION ABSTRACT]