Using quantitative pollen-based land-cover estimations and a spatial CA_Markov model to reconstruct the development of cultural landscape at Rouge, South Estonia

Quantitative pollen-based land-cover reconstruction covering the last 4,000 years was performed using transformation coefficients derived from a modern pollen land-cover database and a palynological record from an annually laminated sequence in Lake Rouge Tougjarv. Proportions of four land-cover classes characteristic of cultural landscape were reconstructed: habitation area, arable land, grassland and woodland. A land-use change model using CA_Markov analysis was applied for spatial reconstructions for seven periods: 600 B.C., A.D. 300, 800, 1200, 1700, 1870 and 1940. Historical maps from A.D. 1684, 1870-1899 and 1935 were used for calibration of quantitative estimates and to validate spatial reconstructions. The accuracy of the estimates depends on the availability of modern analogues and differs among land-cover classes, being highest for classes with directly connectable pollen indicator types (arable land, forest) and lowest for settlement areas. Spatial reconstructions produced by the CA_Markov land-cover change model show moderate accordance with historical data. However, the large uncertainties in land-cover input data must be considered in the evaluation of the KIA results of the spatial model. Permanent low intensity, rural land-use in the Rouge area started at the beginning of the Bronze Age (c. 1800 B.C.). The major increase in the extent of rural land-use took place at the beginning of the 13th century and culminated during the 19th century when c. 90% of the RSAP of Rouge Tougjarv was open. During the last century, rural land-use decreased constantly.