| Study on Forest landscape boundary is an extension of forest boundary ecology on macroscale. In this paper, GIS, RS, GPS and multivariate statistics methods are combined with field investigation. At first, the position and width of boundary between tundra and mountain birch and between mountain birch and evergreen coniferous forest in the northern slope of Changbai Mountain were quantitatively determined. Secondly, the dynamic changes of forest landscape boundary were illustrated. And then, the relationships between forest landscape boundary change and climatic variations were analyzed. Finally, the relationships between the distribution of forest landscape boundary and environmental gradients were demonstrated. Based on transect data and remote sensed data, two methods, moving split-window method and PCA(principal component analysis) were compared. The results show that the width of the boundary between tundra and mountain birch is 60 meters by moving split-window methods. The boundary width between mountain birch and evergreen coniferous forest is 50 meters according to transect data, 60 meters according to remotely sensed data by the same method. However, because of the discontinuous distribution of mountain birch, it is difficult to determine width of the boundary between tundra and mountain birch with PCA method. But width of the boundary between mountain birch and evergreen coniferous forest could be ensured, which is 50 meters based on transect data, 30 meters on remotely sensed data according to PCA. In conclusion, remotely sensed data can be used to determine the width of landscape boundary. Moving split-window method is more stable than PCA. But PCA can also be used to determine the width of landscape boundary when transect lines are properly located. Three Landscape type maps were derived from MSS image in 1975, TM in 1985 and 1997 by supervised classification and interpretation. Index system used for describing landscape boundary was established. On class and landscape scale, dynamic change of forest landscape boundary in Chnagbai Mountain was analyzed in five aspects, length, density, contrast, shape and diversity. The results indicate that forest landscape became more fragmented and landscape shape more complex due to forest cutting, agricultural and other human activities. Regression models of eleven climatic factors against longitude, latitude and altitude were first established. The 11 factors are annual mean temprature, annual precipitation, January mean temprature, July mean temprature, ≥5℃accumulative temprature, ≥5℃ accumulative days, dry index, humid index, warmth index, coldness index, and wetness index. Each model was tested by t distribution, showing that the significance of all these regression models are prominent. That is to say, we can evaluate values of these climatic factors by the models. Using map algebra in the GRID module of ARC/INFO, grid surfaces of the 11 climatic factors were generated. And the surfaces were also created by another method called KRIGING which is a useful command in ARC/INFO. By comparing the surfaces created with these two methods, we found that annual mean temperature and January mean temperature are not suitable to be interpolated by KRIGING. Then, the climatic range of forest landscape types in Changbai Mountain was quantitatively determined. And finally, the relationships between climatic change and landscape boundary variation were analyzed, including tundra, mountain birch and evergreen coniferous forest boundary, Results show that tundra first moves up and then down from 1975 to 1997. The upper elevation limit of mountain birch has the same trend as tundra, but the lower limit moves down. Only the upper elevation limit of evergreen coniferous forest moves up during the same period. However, its lower limit has no rules to follow. The statistical data of ground observation stations demonstrated that these changes might be related to variations of January mean temperature, July mean temperature and annual precipitation. The range of elevation, within which the major landscape types in Changbai Mountain distribute, was measured precisely by using GIS and RS. The optimum elevation range of tundra is from 1780meters to 2212meters, mountain birch 1705 to 1956 meters, evergreen coniferous forest 1042 to 1625 meters, broad leaved and korea pine forest 823 to 1184 meters. We also measured the range of temperature and precipitaion. From tundra to mountain birch, to evergreen coniferous forest, to broad leaved and korea pine forest, the optimum range of temperature is -4.75～-2.40℃,-3.42～-2.07℃,-1.49～1.39℃ and 0.71～2.37℃ respectively; the optimum range of precipitation is 1034mm～1110mm,1014mm～1060mm,883mm～1017mm , 824mm ～ 925mm respectively. At last, the relationships between landscape distribution and slope and aspect were analyzed. The results show that slope and aspect have little effect on the distribution of forest landscapes in Changbai Mountain nature reserve.
|
PDF Full Text Request