臺灣山地雲霧森林種子雨與樹苗之時空動態

Abstract

在樹木更新之長期過程中，樹木種子傳播和樹苗動態影響了森林群落構建，因此對於維持森林物種共存至為重要。但由於被掠食與傳播限制，種子轉換至樹苗存在落差，擴大了樹種補充更新在時空上的變異。其中，種子不一定能到達適合樹苗生存的生育地，造就樹苗分布與環境因子的微弱相關。再者，樹苗和成樹組成間也欠缺一致性，這些落差模糊了樹苗階段的更新棲位在森林群落構建中所扮演的功能。 為解決此問題，首須了解樹苗群聚如何對應微環境中的局部異質性。本研究提出地被植物因同時與樹苗和微環境互動，可做為樹苗棲地的空間指標。以2009年在楠溪森林動態樣區（23°27'40.7' N, 120°54'22.2' E，地處台灣山地雲霧森林）所調查的樹苗與地被植物進行分析，結果顯示地被植物對於解釋樹苗群聚的空間變異貢獻度較大，且能用於區分空間上的更新區塊。不同的更新區塊間，較高草本地被植物對樹苗豐度、歧異度以及和成樹組成關聯帶來不同程度的鄰伴效應。 接續三年（2010-2013）的樹苗動態監測也證實了高草鄰伴可帶來對樹苗的庇護效應，並和樹苗物種特性（如耐蔭性、萌發高度）共同作用，抵抗有蹄類動物植食。這促使了在樹苗建立初期，因有蹄類動物植食之死亡率有高度種間異質性。誠然有蹄類動物植食造成樹苗高死亡率，但並未破壞樹苗物種共存。植食致死率之種間異質性，以及在突增的補充更新之後而來的密度依變致死，都抑制了優勢樹苗豐度。我藉由動態模型，也發現更新補充擾動（亦即加入了時間異質性），可使樹苗稀少種避開優勢種樹苗引發的外顯競爭。 以小苗樣點附近的種子網進行五年的種子雨收集（2008–2009, 2010–2012），研究更新補充動態中的時間異質性與其對種子轉換至樹苗間的影響。多數優勢樹種每年週期性地繁殖，也相關於氣候上的季節性，但產出的種子和樹苗補充不論在時空或豐度上都產生不一致。分析各物種在各樣點年間種子－樹苗轉換和各生物與非生物因子的關聯，其具影響力之因子（如傳播前種子損失率、傳播限制、土壤種子庫補充等）在不同種子傳播方式的物種間影響各異。這些研究顯示了在與動物族群與鄰伴植物間的生物交互作用下，如何在樹種種子與樹苗所存在的動態本質間扮演連結的角色，其重要性可提升在森林群落構建中的物種共存。

Seed dispersal and seedling dynamics are crucial for tree species in their long-term regeneration processes by affecting community assembly of a forest. However, the discordance in seed-to-seedling transition, which is due to loss by predation and dispersal limitation, constitutes a great spatiotemporal variation in recruitment of tree species. Seeds are not guaranteed to be able to disperse to the microhabitats favorable to seedling establishment or survival; this results in weak association between seedling distributions and environmental factors. Moreover, the uncoupling between seedling and adult tree assemblages further obscures our understanding of the role of ecological niches specific to the seedling stage (regeneration niches) in forest community assembly. To address the discrepancies at the seedling stage, detecting local heterogeneity in microenvironments, to which seedling assemblages respond, is a prerequisite. I propose that understory plants, which may interact with microenvironments and seedlings at local scale, can be a better spatial descriptor for regeneration habitats. To test this, a demographic survey for tree seedlings and understory plant assemblages was conducted in 2009 in the Nanhsi forest dynamics plot (23° 27'40.7' N, 120° 54'22.2' E), a montane cloud forest of Taiwan. I find that the understory plant spatial structure contributes most in explaining spatial variations of the seedling assemblages and facilitates the identification of patches of different regeneration habitats. Among these regeneration patches, the neighboring effects from tall herbs play an important role in affecting seedling density, diversity, and its coupling with conspecific adult trees. A consecutive seedling dynamics survey for three years (2010–2013) provides further supporting evidences for the neighboring effects from tall herbs. The nurse-plant effect by tall herbs can co-operate with seedling species traits (shade-tolerance, and seedling initial height) against ungulate herbivory, resulting in uneven species-specific mortality at an early stage of seedling establishment. Ungulate herbivory greatly caused seedling mortality here; however, it was not really detrimental to seedling species coexistence. The uneven species-specific mortality by herbivory and density-dependent mortality after bursts of recruitments suppress the abundant seedling species. Furthermore, by dynamic modelling, I also find that fluctuating recruitment, which increases temporal heterogeneity in recruitment dynamics, may prevent rare species from apparent competition from the abundant species. The temporal heterogeneity in recruitment dynamics and the consequence of seed-to-seedling transition are studied using seed rains that had been collected in seed traps nearby the seedling sites for five years (2008–2009, 2010–2012). Most of the abundant tree species reproduced annually and periodically, correlating to the climatic seasonality. But the seed arrival uncoupled with recruitments in space, and in time, as well as in the level of species abundance. By analyzing the specific seed-to-seedling transition across years in each site associated with various biotic and abiotic factors, I find that the influential factors (such as the pre-dispersal loss of seeds, dispersal limitation, seed-bank buffering, etc.) had different impacts among tree species with different seed dispersers. My studies highlight the importance of bio-interactions with animals and neighboring plants that can interplay with the dynamic nature of seeds and seedlings from the dispersal, recruitment, to establishment stage, and thereby facilitate species coexistence in the forest community.