宜蘭地區台灣肖楠人工林林分表直徑分布生長模式

INCORPORATING DIAMETER DISTRIBUTION IN A STAND TABLE PROJECTION GROWTH MODEL FOR CALOCEDURS FORMOSANA PLANTATIONS IN YILAN

汪大雄(Dar-Hsiung Wang) ； 羅仁甫(Ren-Fu Lo) ； 張妙華(Miao-Hua Zhang) ； 古宇任(Yu-Ren Gu) ； 葉晉宇(Jin-Yu Ye) ； 王兆桓(Chao-Huan Wang)

《中華林學季刊》 50卷2期 (2017/06) Pp. 159-172

摘要

森林生長模式之種類繁多，範圍甚廣，其使用之分布範圍可從整個生態系、林分到個別單木，其模式建立之目的可從林分收穫之預測到生態過程之描述。林分表直徑分布生長模式(Incorporating diameter distribution in stand table projection growth model)是從林分直徑級之觀點，直接預測各直徑級內林木在預測期間內之生長變化，以進行未來林分表之更新。本研究是在國立宜蘭大學延文林場25年生之台灣肖楠人工林內，由設置的21個樣區中選取同一撫育處理的6個樣區，面積各為0.04公頃，分別在2005, 2010和2015年，進行生長調查；探討在林分表法(Stand table projection)的生長模式內，使用由Nepal and Somers (1992)提出之NS方法與由Cao and Baldwin (1999b)、Cao (2007)、Allen et al. (2011)所發展之CB方法，此兩種合併直徑分布的推估方法，在不同生長預測時期(5年和10年)下對林分各直徑級生長預測之差異，並和直徑級機率模式比較。研究結果顯示，使用之兩種方法，都能有效地預測林分直徑級內株數，因林木生長導致株數分布之變化。生長預測時期為5年時，所有樣區均通過K-S一致性檢定(α=0.05)；而生長預測時期為10年，CB法有1個樣區未通過K-S一致性檢定。兩種方法均顯示預測期間10年之推定誤差，較短期之5年之推定誤差為高，表示此兩種修正的林分表法用在短期預測之效果較佳。此外，在描述林木直徑級株數分布時，林分表直徑分布生長模式較林分直徑級機率模式，更能有效地描述株數在直徑級之細微變化。

關鍵字

林分直徑分布 ； 林分生長模式 ； 林分表預測

並列摘要

With a variety of application, forest growth models are used widely from stands to ecosystem to describe the process of ecosystems or predict the growth of stands and individual trees. Focusing on the stand diameter class viewpoint, the growth models of stand table projection incorporated with diameter distribution are used to directly predict the change of frequency distribution in diameter class due to tree growth occurring in the growth periods. In this study, we set up 6 plots of 0.04 ha each in the Calocedurs formosana plantation on the National ILan University Experimental Forest in Yilan, and carried out inventories in 2005, 2010, and 2015, respectively. Two approaches respectively proposed by Nepal and Somers (NS), Cao and Baldwin (CB), were used to compare efficiency on the prediction of the change in stand table for two periods of 5 years and 10 years, and finally, we compared stand diameter distribution models in depicting the tree frequency distribution among diameter classes. Results show that both methods can be used to efficiently predict the change of tree frequency in the diameter classes due to tree growth. The K-S test, used to check the consistency in trees frequency in stand table for both methods, demonstrate their validity for all plots investigated, except for a 10-year growth projection in one plot of the CB method (α=0.05). The less prediction error in the 5-year growth period for both methods indicates that stand diameter class growth models are more suitable for the short term prediction. Moreover, this study shows that the stand diameter class growth models are more capable of describing the real trend in diameter among the diameter classes than stand diameter distribution model do.

並列關鍵字

Stand diameter distribution ； Stand growth model ； Stand table projection