Abstract:

Based on the Landsat TM/ETM remote sensing image data of 1990 and 2013 of the Naoli River Valley in the heart of the Sanjiang Plain, the land use dynamics model and land use intensity model were used in combination with the Markov transfer matrix to analyze changes in land use during the period of 1990-2013 and the CA-Markov model was to simulate and predict land use pattern in the Naoli River Valley in 2025. Results show that the changes in land use structure was mainly dominated by conversion of farmland from paddy to upland or vice versa and alienation of farmlands and reclamation of wastelands. The conversion of upland to paddy was the main feature of the changes in landscape. Reclamation of unused land led to substantial reduction of wasteland in area. The land use intensity index of the valley rose from 251.46 in 1990 to 270.48 in 2013, which indicates that land use increased steadily in exploitation degree in the Naoli River Valley. The simulation using the models demonstrates that the Kappa coefficient of the valley was 0.82, indicating that the prediction was reliable. The variation of land use types in the valley in 2013-2025 may differ from that in 1990-2013 in trend and rate. The change in woodlands will be the widest in range and the area of paddy fields will keep on rising, but at a declining rate with the single dynamic degree falling down to 2%; The area of upland will stay almost unchanged. The land use intensity index will continue rising and the valley will continue to be disturbed by human activities with rising intensity during the period from 2013 to 2025.

Key words: CA-Markov model, Naoli River Valley, LUCC, spatial analysis model