民墾地之地景變遷監測研究

Monitoring of Landscape Change on Unlawful Cultivated Forestland

吳治達(Chih-Da Wu)

指導教授 ： 鄭祈全

國立臺灣大學/生物資源暨農學院/森林學研究所/碩士(2004年)

https://doi.org/10.6342/NTU.2004.00502

摘要

摘　　要

森林地景變遷監測之目的在掌握過去及現在地景結構的變化過程，並進一步推估未來演變趨勢，以提供森林經營管理之依據。相關研究目前已成為國內外林業科學之重點方向。本研究利用與人為干擾有關之八項量化地景結構指數，比較蓮華池研究中心民國12年、60年及87年三期之地景結構變遷情形，並以Shannon多樣性指數t檢定法評估各期間民墾干擾程度之差異，最後整合馬可夫鏈模式與二項式logit廻歸之分析結果，預測未來之地景變遷和模擬民墾地發生之空間分佈情形，並進而探討民墾地干擾對未來地景之影響效應，提供蓮華池研究中心未來經營管理之參考。本研究結果如下：
1. 三期地景結構分析之結果顯示，由於受到民墾地干擾的影響，造成區塊總數、區塊密度、邊緣密度、形狀指數、碎形維度、Shannon多樣性及Shannon均勻度指數值上升，而平均區塊面積指數值下降之趨勢，惟就Shannon多樣性指數之t測驗檢定結果顯示，雖然三期的Shannon多樣性指數有差異，但其地景變遷在5%顯著水準時，均未達顯著差異，此說明了該期間之民墾地干擾並未達顯著的差異程度。
2. 使用馬可夫鏈模式預測未來之地景變遷結果顯示，根據民國60年至87年間之資料所建立的馬可夫鏈模式，經模式確認與驗證後，適合應用於預測民國141年、195年及249年之地景結構，並發現林業用地面積從民國60年的88.36%和民國87年的85.87%，降為民國141年的81.41%、民國195年的77.18%及民國249年的73.17%；而民墾地面積則由原來的0.90%和3.39%依序增加為7.85%、12.08%及16.09%。至於應用二項式logit廻歸分析模擬民墾地發生之空間分佈結果顯示，民墾地的發生與海拔、坡度、距民墾地、道路及河流的距離均有密切的關係，但其中以距民墾地和距道路的遠近最有相關。
3. 在探討民墾地干擾對未來地景之影響效應方面顯示，民國141年、195年及249年之Shannon多樣性指數隨著民墾地面積之增加而增加，分別為0.61，0.70和0.76，同時，由Shannon多樣性指數之t測驗檢定結果顯示，雖然到民國141年仍未發生民墾地干擾的情形，但到了民國195年，其干擾程度則已達5%顯著水準，而到了民國249年甚至達到1%的極顯著水準。
綜合以上結果可知，整合馬可夫鏈模式和二項式logit廻歸分析預測未來之地景變遷和模擬民墾地發生之空間分佈情形，確實為一有效、可行的方法，此外，Shannon多樣性指數之t測驗檢定亦可用以監測和評估民墾地干擾對未來地景之影響效應，因此，本研究所得結果可提供蓮華池研究中心經營管理之參考。

關鍵字

地景結構 ； 二項式logit廻歸分析 ； 馬可夫鏈模式 ； 地景變遷監測 ； 空間分佈 ； 民墾地

並列摘要

ABSTRACT

The purpose of monitoring forest landscape change is not only to understand the change of past and present landscape structure, but also to predict the future trend for providing the reference of forest management. Therefore, the related research has been growing concern by forester. This study applied eight qualified landscape structure indices, which are related to human disturbance, to compare the landscape change of Lien-Hwa-Chih Research Center among 1923, 1971, and 1998. The t-test based on Shannon diversity index was also used to examine the disturbance differences caused by unlawful cultivated forestland. In addition, the mathematical models such as Markov chain and binomial logit regression analysis were applied to predict the future overall landscape structure and to simulate the spatial distribution of unlawful cultivated forestland. Again, the t-test was applied to monitor the future effect of disturbance caused by unlawful cultivated forestland. The results are as follows.
1. The comparison of landscape structure among three different dates indicated that patch number, patch density, edge density, shape index, fractal dimension, Shannon diversity index, and Shannon evenness index were increased due to the disturbance effect of unlawful cultivated forestland. Conversely, the mean patch was decreased. As for the t-test examination using Shannon diversity index, the landscape change among 1923, 1971, and 1998 showed no differences at the 5% significance level, that is, there was no significant differences about the disturbance caused by unlawful cultivated forestland.
2. The prediction of future landscape change using the Markov chain model indicated that the Markov chain developed by the time step (1971-1998) was matched well after model verification and validation. The projection of overall landscape structure from 1971 to 2052, 2106, and 2160, respectively showed that forestland was decreased from 88.36% in 1971 and 85.87% in 1998 to 81.41% in 2052, 77.18% in 2106, and 73.17% in 2160. The unlawful cultivated forestland was increased from 0.90% and 3.39% to 7.85%, 12.08%, and 6.09%, respectively. As for the spatial distribution using binomial logit regression analysis, the result showed that the occurrence of unlawful cultivated forestland had the relationship with elevation, slope, and the distance of road, river, and previous unlawful cultivated forestland. However, it was strongly correlated with the distance of road and previous unlawful cultivated forestland, and less correlated with elevation, slope, and the distance of river.
3. On the aspect of monitoring the future disturbance effect caused by unlawful cultivated forestland, the Shannon diversity indices among three predicted time steps were increased. They are 0.61, 0.70, and 0.76, respectively. Meanwhile, the Shannon t-test indicated that although there’s no disturbance effect in 2052, it would occur at the 5% significance level in 2106 and had the significant effect of disturbance at the 1% significance level in 2160.

From the result, obviously the integration of Markov chain and binomial logit regression analysis is a feasible approach to predict the overall landscape change in future and to simulate the spatial distribution of unlawful cultivated forestland. Also, the Shannon t-test is a good way to monitor and assess the disturbance effect of unlawful cultivated forestland on future landscape. The above result can be provided to Lien-Hwa-Chih Research Center for the reference of forest management in order to prevent the occurrence of unlawful
cultivated forestland.

並列關鍵字

Unlawful cultivated forestland ； landscape structur

參考文獻

肖篤寧（1993）景觀生態學－理論、方法及應用。地景企業股份有限公司出版　臺北市。

吳守從、陳永寬（2002）結合航測資訊與地景指數探討南仁山生態保護區之干擾

野葙F、鄭祈全（2000）應用馬可夫模式探討六龜生態系經營區之地景變遷。臺灣林業科學　15(1):41-49。

野葙F、鄭祈全（2002）應用GIS及Logit模式分析六龜地區林地變遷。航測及遙測學刊　7(2):61-74。

鄭祈全（1999）森林地景變遷之監測研究。臺灣林業科學　14(4):493-507。