日本丹澤山柳杉林之樹液流特性

Abstract

蒸散是森林水循環中主要的一部分。樹液流測量法是普遍用來測量個體和林分蒸散的方法。雖然有許多前人文獻提到個體樹的樹液流有明顯的空間和時間變異，卻甚少有研究檢測其變化的季節性。為了瞭解蒸散的機制，釐清樹液流的空間變異及時間變異是必須的。因此，本研究的目的為 1) 檢測樹液流的方向變異及其季節性 ; 2) 檢測樹液流和氣象因子(蒸氣壓差及太陽輻射)之間的時間差及其季節性。目的1和2的研究地點位於日本神奈川縣西部的丹澤山柳杉林，利用32組Granier探針安裝置16棵樹，每棵樹各安裝兩個方位來測量其樹液流。資料收集的期間是從2017年8月21日至2018年4月30日; 3) 比較位於溫帶氣候，且氣候季節性較明顯的日本丹澤山，及位於亞熱帶氣候，且氣候季節性較不明顯的臺灣中部，溪頭柳杉林兩者的分析結果。 首先，根據結果顯示樹液流的方向變異有季節變異。秋天到冬天，西南方的樹液流增加，東北方的樹液流卻有減少的趨勢;相反的，冬天到春天時，西南方的樹液流減少，東北方的樹液流卻有增加的趨勢。造成方向變異之季節變異的原因可能和太陽照射樹冠比例的季節變異有關。其次，樹液流和氣象因子之間的時間差也有明顯的季節變異。在秋天，樹液流和蒸氣壓差之間的平均時間差為122分鐘，和輻射之間的平均時間差為163分鐘；在冬天，樹液流和蒸氣壓差的平均時間差為188分鐘，和輻射之間的平均時間差為214分鐘。另一方面，在夏天，樹液流和蒸氣壓差的平均時間差為54分鐘，和輻射之間的平均時間差為116分鐘。綜合上述，可得知在秋冬的平均時間差較夏天長。土壤乾旱並不是造成較長時間差的原因，推測可能有其他原因，像是枝條附近的微氣候，植物體內維管系統的剖面結構等。最後，台灣和日本兩者柳杉林比較的結果，相較之下，樹液流和蒸散之間時間差的季節變異在台灣溪頭柳杉林較不明顯。在溪頭，方向變異模式的不一致和時間差的季節變異較不明顯的原因，可能是因為其樹液流的空間變異和時間變異的季節性較不明顯。此研究結果顯示，即使是同一物種，其樹液流的特性在溫帶和亞熱帶氣候仍有所不同。因此，利用樹液流測量法推估蒸散時，要將地點及季節等因素列入考量，設計出適合當地及不同季節的探針安裝方式。

Transpiration is a major part of the water cycle in forested ecosystems. Sap flux (F_d) measurement is a well-known technique for investigating transpiration at the individual-tree and forested-stand levels. Although numerous studies have indicated significant spatial and temporal variations among the F_d values of individual trees, few have examined seasonal patterns corresponding to these values across multiple trees. Spatial and temporal variations in F_d and related seasonal affects must be determined to enable estimations of individual and stand transpiration as well as clarification of the transpiration mechanism. To this end, the present study examined 1) azimuthal variations in F_d and the related effects of seasonal changes, and 2) time lags between F_d and meteorological factors such as vapor pressure deficit (VPD) and solar radiation (SR), as well as the related effects of seasonal changes. These investigations were conducted in a Japanese cedar forest in the Tanzawa Mountains, located in the western part of Kanagawa Prefecture, Japan under a humid temperate climate. Additionally, 3) to characterize the results derived from the samples in the Tanzawa Mountains, the data sets were compared with those from previous measurements in a Japanese cedar stand in Xitou, located in central Taiwan. The Xitou stand is characterized by a humid subtropical climate with less distinct seasonality than that of the temperate climate in the Tanzawa Mountains. For objectives 1 and 2, 32 sets of Granier-type sensors were placed on 16 trees to measure F_d in two directions for each tree. Data were collected from August 21, 2017 to April 30,2018. The main results are summarized as follows. First, azimuthal variations in the F_d of individual trees corresponded with seasonal changes. From autumn to winter, F_d on the southwest sides increased, but F_d on the northeast sides decreased. From winter to spring, however, F_d on the southwest sides decreased, but F_d on the northeast sides increased. These seasonal azimuthal variations may have been related to seasonal changes in crown exposure to sunlight. Second, the time lags between F_d and the meteorological factors also clearly exhibited seasonal changes: the time lags were longer in autumn (122 and 163 min for VPD and SR, respectively) and winter (188 and 214 min for VPD and SR, respectively) than in summer (54 and 116 min for VPD and SR, respectively). Soil drought was not the cause of the longer time lags. Possible causes included the microclimate around branches and the anatomical characteristics of the vascular systems. Third, comparison of the results for the Tanzawa Mountains with those for Xitou in Taiwan revealed that seasonal changes in time lags were not as pronounced in Xitou as in Tanzawa. The inconsistent temporal patterns of azimuthal variation in F_d and indistinct seasonal relationship with time lags in Taiwan may be attributed to the less distinct seasonality of the spatial and temporal variations in F_d in Taiwan compared with Japan. This study indicated that the characteristics of F_d differed between temperate and subtropical climates, even for the same species. Therefore, different strategies should be adopted according to season and region in estimates of transpiration based on F_d.