The temporal and spatial pattern of vegetation change in the transition from estuary to freshwater swamp: Whangapoua Estuary, Great Barrier Island, northern New Zealand

Abstract

This study examines vegetation dynamics during the transition from salt marsh to freshwater swamp (hydrarch succession) at Whangapoua Estuary, Great Barrier Island. Sediment cores from eleven sites ranging from mangrove through salt meadow to freshwater swamp were analysed for pollen, charcoal and sedimentation rates in order to reconstruct a c. 3500 year record of vegetation change. Studies of vegetation communities, modern pollen rain and a comparison of the results with mid Holocene palynological investigations from Great Barrier Island were undertaken to enhance the understanding of spatial and temporal aspects of vegetation succession in wetlands of this area. There are six broad vegetation communities present within Whangapoua Estuary. (A) Mangrove (Avicennia marina) mudflats, (B) Juncus kraussii sea rush, (C) Oioi (Leptocarpus similis) salt meadow, (D) Baumea juncea sedges, (E) Manuka (Leptospermum scoparium) shrubland, (F) Raupo (Typha orientalis)/Cabbage tree (Cordyline australis) swamp forest. The pattern of species distribution from salt marsh to freshwater results from the interaction between species and the physical constraints of the salinity/inundation gradient and the freshwater inundation gradient. This pattern accounts for 77% of the floristic variation in the modern vegetation, indicated by the different vegetation communities exhibiting similar distribution patterns along the estuarine to freshwater gradient. The quantitative relationships between pollen representation and source vegetation frequency varied considerably between species. Pollen assemblages at the seaward end of the salinity gradient are less clearly representative of the associated vegetation than those at the landward end. This is probably because the open vegetation at the seaward end allows the influx of many wind- and water-dispersed pollen types. However, if the long-distant and over-represented pollen rain are excluded, five of six of the broad communities are represented by their pollen spectra. The exception is the Juncus kraussii community. This is enhanced by Boxplot analysis which indicates that the pollen proportion of Avicennia (mangrove), Leptocarpus similis (salt meadow), Plagianthus divaricatus (sedges), and Cordyline (swamp forest) are highly discriminatory variables in relation to vegetation type. This study added new information for the pollen representation of several salt-marsh species, such as Leptocarpus similis (well-represented), and Selliera and Triglochin (under-represented). Juncus kraussii is non-represented in pollen spectra. The fossil pollen result indicates that three major zones can be distinguished in seven of the eleven cores, while the other four cores also agree with pollen trends for the upper two zones. The lower zone (Pre-Impact, c. 3500-750 cal. yr B.P) is characterised by a phase of marine sedimentation. The local environment at c. 3500 cal. yr BP was a tidal flat surrounded by conifer-hardwood forest (Dacrydium, Libocedrus, Prumnopitys and Metrosideros) with Cyathea tree ferns, indicating a moist warm climate. After c. 1700 cal. yr BP, Agathis and Phyllocladus became more common, and more shrubs such as Myrsine appeared in the sub-canopy, suggesting the climate was becoming drier. Around 1500 cal. yr BP, the appearance of Avicennia pollen marks the start of the successional vegetation sequence, and coincides with charcoal fragments, probably caused by natural fire. The intermediate zone [Polynesian, c. 750 cal. yr BP (c. 1200 AD) to 110 cal. yr BP (1840 AD)] encompasses large human impact in the region and the transition of the core sites from mainly marine to mainly brackish. The associated decline in tree pollen, a result of deforestation by fire, coincides with a sharp and sustained increase in charcoal and Pteridium spores. Higher sedimentation rates at the same time indicate increased erosion of the surrounding hills. Around 350 cal. yr BP (c. 1600 AD), a sharp decrease in Leptocarpus similis pollen and an increase in Baumea pollen indicate the transition from salt marsh to brackish swamp in several cores. The uppermost zone (European, 1840 AD to present) is marked by the appearance of Pinus pollen and an increase in Poaceae pollen and greater freshwater input. A local change from Baumea to Leptospermum/Gleichenia and then to Typha swamp is coincident with continued deforestation during this period and a further increase in the rate of erosion. The driving factor for plant succession in the Whangapoua Estuary during the late Holocene thus appears to have been siltation caused mainly by human impact, and this infilling may also have been enhanced by sea-level recession c. 700 yr BP. The pollen sequences from 18 profiles in Great Barrier Island indicate that the early marine succession route is linear, and is primarily under the control of allogenic factors. The shift between allogenic and autogenic controlling factors is represented by the appearance of Baumea, which is the starting point for a greater variety of vegetation succession routes. The freshwater succession routes are multiple and are associated with an increase in autogenic factors and greater variability of allogenic processes in the later successional stages. The spatial vegetation pattern actually reflects the changes which have occurred through time in salt marsh. There are some variations between spatial and temporal vegetation in freshwater wetland because there are increased varieties of successional pathways once the system escapes the tidal regime. There is no natural 'climax' stage in the hydrosere process. Disturbances (natural and human), interacting with autogenic factors (peat-forming plants), have been driving factors of vegetation succession from estuarine to freshwater wetland on Great Barrier Island ever since sea levels stabilized c. 6500 yr BP. However, the results from Whangapoua Estuary show that human impact was the most powerful factor. The results lead to a different view from the classical linear model of succession, and support a more dynamic concept in which some areas may proceed through time to forest, while other areas (depending on the mix of auto- and allogenic factors) may remain as organic wetlands.