Introduction

Millions of birds undertake annual long-distance seasonal migrations along the East Asian-Australasian Flyway (EAAF), which stretches for over 13,000 km from the Arctic Circle to Australia and New Zealand [1]. The ecological functionality of this migration corridor depends on a network of coastal wetlands that represent steppingstones used by the birds as staging areas and over-wintering sites [2, 3]. Over the past several decades many of the wetlands along the EAAF have been converted for development, from agriculture and aquaculture to coastal infrastructure [4–7]. Over 65% of the intertidal mudflats in the Yellow Sea, one of the more important staging areas along the flyway, have been lost in recent decades [8–10]. Continued loss of strategically located wetlands could drastically impact connectivity, preventing birds from completing their migrations, resulting in eventual species population declines, extirpation and eventual collapse of this important ecological and evolutionary phenomenon that has evolved over millennia [2, 3, 11].

But like many coastal wetlands across the world the EAAF wetlands are also under additional threat from global climate change and its consequences, notably sea level rise (SLR) and extreme storms that are accompanied by stronger wave and tidal surges [5, 12]. The impacts of SLR and natural disasters are further compounded by subsidence and decreasing coastal sedimentation that is necessary to counter SLR [13–16]. Iwamura et al. [17] estimate that between 23 to 40% of the intertidal habitat will be lost and unavailable for migratory birds along the EAAF because of sea level rise, with some regions losing more than others. Coastal development, which has resulted in loss of wetlands, can now also place the remaining wetlands in a situation where they will be subjected to coastal squeeze; unable to migrate further inland in the event of SLR because of the adjacent hard infrastructure [18–20]. Thus, important wetlands along the flyway should be assessed for the impacts of climate change and associated drivers of ecological and environmental change to plan and implement proactive climate adaptation strategies. These analyses should be conducted at site scales to identify site-scale conservation strategies based on habitat availability, current and planned landuse, and restoration opportunities. Because wetland conversion and coastal development is still continuing apace these analyses are urgently needed before options for conservation and restoration disappear.

In Hong Kong, projections suggest rising sea levels and warmer, wetter weather that can spawn a higher number of severe typhoons [21, 22]. But, while the sea is rising, the deltas and mudflats in the Pearl River Delta could be shrinking. A model of sediment deposition and coastal nourishment of the Pearl River Delta projected that the fluvial sediment supply that fed the Pearl River Delta has decreased by 71% from the previous ‘natural state’, and any potential compensation from increased river flows attributed to climate change will only be about 1% of this deficit by the end of the 21^st century, assuming impacts from human activities are maintained at current levels [23]. Under this scenario, the coastal zones will be starved of approximately 6,000 Mt of fluvial sediment until the end of this century, which may also result in the erosion and shrinking of the delta, including the inter-tidal mudflat [23]. In fact, recent studies of the sub-aqueous bathymetry of the Pearl River Delta has shown that sections of the delta are already eroding, or accreting at lower rates [24, 25].

The Mai Po Inner Deep Bay Ramsar site in Hong Kong is an important staging and wintering site along the EAAF and supports large numbers of migratory birds, including 24 globally threatened bird species associated with wetlands. Thus, rising sea levels, stronger storm surges and shrinking mudflats will combine to decrease the availability of habitats for the waterbirds in Mai Po in the future, and adaptation plans to counter the threats from climate change and offshore subsidence are needed now to ensure persistence of migrations and flyway functionality. Here, we modelled the impacts of climate change and decreasing sedimentation rates on the important bird habitats in the Mai Po Inner Deep Bay Ramsar site to support a climate adaptation strategy that will continue to host migratory waterbirds. To do this, we applied the SLAMM (Sea Level Affecting Marshes Model) version 6.7 [26] to simulate habitat conversion under selected sea-level rise scenarios for coastal areas in the Mai Po Inner Deep Bay Ramsar site and adjacent areas surrounding it.

This model has been widely used for coastal resource management across the world, including for selected wetlands in the Yellow Sea estuary of China [16, 27] and Korea [28] in the EAAF region. But we note that the lack of data on historic wetland conversion, especially outside North America [29] hinders retrospective analysis, which could be used to calibrate and validate the model. Analyses have shown that elevation data is the most important factor for output accuracy since conversion among habitat classes is mainly governed by elevation [29–31]. Therefore, a high-resolution DEM is important for generating consistent results. Other site-specific parameters such as tidal range and accretion also affect the model outputs and should therefore be as accurate as possible [30]. Despite these shortcomings, SLAMM is the only marsh migration model that handles uncertainty in spatial inputs (DEM, VDATUM, etc.) and parameter choices, and this capability has contributed to its wide use [31].

Mai Po Inner Deep Bay Ramsar site

The Mai Po Inner Deep Bay Ramsar site (hereafter Mai Po) is located in the North West New Territories of Hong Kong Special Administrative Region (HKSAR), along the southern edges of the Deep Bay of the Pearl River estuary (Fig 1A). The Ramsar site is comprised of tidal mudflats, coastal mangroves, marshes, and fishponds (Fig 1B), and now represents Hong Kong’s largest remaining wetland. It is about 1,500 ha in extent, with a 380 ha nature reserve zoned within it (Fig 1B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1.

The location of the Mai Po Inner Deep Bay Ramsar site in relation to the Deep Bay (A, above) and major habitat types (B, below) that were cross-walked to the SLAMM. categories, as indicated in Table 1. The Shenzhen River borders the northern boundary of the Wetland Conservation Area. The Kam Tin River is south of the Mai Po SSSI. Basemap Sources: Esri, DeLorme, HERE. Landuse maps and boundary layers for the wetland conservation area, Ramsar site, and Nature Reserve are from WWF Hong Kong. https://globil-panda.opendata.arcgis.com/datasets/mai-po-wetland-habitats.

https://doi.org/10.1371/journal.pone.0239945.g001

Mai Po supports about 80,000 wintering waterbirds and another 20,000–30,000 use it as a staging area during their annual migration in spring and autumn. Regular visitors to Mai Po include the Endangered Black-faced Spoonbill (Platalea minor) and Nordmann’s Greenshank (Tringa guttifer), and, occasionally, the Critically Endangered Spoon-billed sandpiper (Calidris pygmaea). Mai Po played a significant role in the population recovery of the Black-faced spoonbill, and now regularly supports about 7–10% of the global population [32].

In the 1950s, most of the mangroves in Mai Po and surrounding coastal areas were converted to a system of traditional aquaculture pond management known as gei wai, where the ponds were connected to the bay via a sluice gate that allows water exchange with the bay. Some mangroves were left to remain in the ponds, and the biomass from the mangroves nourished the shrimp, oysters and fishes that were stocked in the ponds through tidal water exchanges, which also aerated and regularly flushed the ponds. Shrimp were harvested in the summer, and the gei wai were drained during the Autumn-Winter for harvesting bigger fishes and maintenance. During this time the birds are able to use them as feeding areas. Thus, these gei wai are a good example of ‘wise use’ management of wetlands that supported both migratory waterbirds and local fisher livelihoods. Since the 1960’s, however, most gei wai have been converted to non-traditional fishponds with no connections to the bay and used for commercial grow out of fishes, mostly exotic species. However, the gei wai in the Mai Po Nature Reserve are operated in the traditional way to maintain habitat for waterbirds.

The nature reserve is also part of the Site of Special Scientific Interest (SSSI), zoned because of its biological importance, and is comprised of seven Biodiversity Management Zones managed under controlled hydrological regimes to provide shallow and deep open water areas, reedbed dominated ponds, rain-fed ponds, and gei wai to provide the range of habitats and food for waterbirds and other birds that use the nature reserve. Several active, commercial fishponds surround the nature reserve, inside the landward areas of Mai Po and beyond, in an area designated as a Wetland Conservation Area (Fig 1B).

Methods

We projected impacts for 3 time intervals, 2050, 2075, and 2100 in relation to 2019 for two different SLR scenarios, 1.5 m and 2.0 m until 2100, and for four accretion rates; i.e., 2mm/yr, 4 mm/yr, 8 mm/yr, 15 mm/yr. We used these accretion rates because historical records (between 1898 and 1949) estimate accretion at 8 mm/year, but the rate had increased to about 15 mm/yr in the 1980’s and 1990’s, attributed to land clearing for agriculture and development in Shenzhen [23, 33]. From the 1990’s onwards, there has been a decrease in fluvial sediment supply to the coast, and models suggest that sediment loads are expected to decrease further, by about 70% in the future with sediment trapping behind the Yangtan and Longtan mega dams [23]. Therefore, we applied 2 and 4 mm/yr accretion rates, which are approximately 30% of the 8 and 15 mm/yr rates, assuming a proportionate deposition based on declining sediment loads.

Models of SLR [21] projected an average rise of about 0.67 m and 0.84 m for Representative Concentration Pathways (RCP) 4.5 and 8.5 respectively, in the seas off Hong Kong, which is 0.2m higher than the global mean values projected by the Fifth Assessment Report of Intergovernmental Panel on Climate Change [34]. Global temperature increases between 1.9 and 3.1°C can bring about fast Antarctic disintegration that can contribute an additional 0.21 to 0.74 m to SLR this century within 5 and 95% quantiles, respectively, of the RCP 8.5, and contributions of up to 0.45 m for the 95% quantile under RCP 4.5 [35]. The temperature range includes the 2°C target set by the Paris CoP. Thus, in this analysis we used 1.5 and 2 m SLR scenarios to approximate the 1.5 m (median) and 2 m (95% quantile) for RCP 8.5 [35], because experts believe that even the upper bound of the ‘business as usual’ RCP 8.5 trajectory could likely be exceeded as few countries are adhering to their Nationally Determined Contributions to mitigate global climate change [36]. This justifies the use of SLR scenarios that even exceed 2 m SLR for future casting and planning [37].

To execute the SLAMM, we first reclassified the habitat and land use-land cover in the Ramsar site and the Wetland Conservation Area (Fig 1) into corresponding SLAMM categories (Table 1). We used expert opinion based on our experience in the Mai Po area and ecological definitions of the SLAMM categories in Clough et al. [26] to do the cross-walk. We used a 5m Digital Elevation Model for Hong Kong to calculate percent slope using the function in ESRI ArcMap 10.7. These three layers (i.e., land cover, DEM, and slope) were used as the main input files for the simulations. Since there was no information on dikes this option was not applied, along with the soil saturation parameter. We also used the following input parameters: offshore direction is west; MTL-NAVD88 (m) = 0 and 0.7m, for low and high tide, respectively; GT Great Diurnal Tide Range (m) = 1.46 m, based on information from the Hong Kong Observatory data (https://www.hko.gov.hk/en/tide/predtide.htm?s=TBT).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Land use and habitat categories in Mai Po Inner Deep Bay Ramsar site and Wetland Conservation Area cross-walked to SLAMM categories (based on SLAMM 6.7 technical documentation, Clough et al. [26]).

https://doi.org/10.1371/journal.pone.0239945.t001

In this analysis, low and high tides are used to denominate different modelling scenarios. Since the average tidal range for the site is 1.46 m, we applied two scenarios of simulation; low and high tide. For the low tide scenario, we simulated no height difference between elevation data and local datum. For the high tide scenario, we applied 0.7 m of height difference, which represents half of the total tidal range. The objective was to derive two results to define a range (i.e. minimum and maximum), since the exact delineation of the shoreline is unknown.

The choice for 0.7 m was made after a sensitivity analysis over the “time-zero” run of SLAMM to calibrate the model to site-specific data. We then tested the model with increasing heights until we perceived changes in the classes and selected 0.7 m as the upper boundary, therefore called high-tide scenario.

Results

There was no discernible impact to the habitats in Mai Po until after 2075, and we do not present the results for the earlier timelines here. However, projections for 2100 suggest that overall, the mangroves, marshes and tidal flats are impacted in almost all scenarios of SLR and accretion (Table 2). Under a 1.5 m SLR scenario, at low tide there was little change in the tidal mudflats and coastal mangroves if the current accretion levels of 8 mm/yr or more continues into the future (Table 2, Fig 2). However, if accretion levels decrease, as predicted, to 4 mm/yr, the tidal flats are expected to be inundated even during low tide with over 77% being lost as habitat for wading and shorebirds, while the mangroves remain relatively unaffected (Table 2, Fig 2). When accretion levels decline to 2 mm/yr, over half the coastal mangroves could also be lost due to permanent inundation of the coastal areas, including during low tide (Table 2, Fig 2). A 1.5 m SLR will also result in a loss of 15 to 18% of inland and open water marshes in the northern reaches of the wetland conservation area (Table 2, Fig 2). These marshes represent the fishponds, and the area bordering the lower reaches of the Shenzhen River and upper estuary, where the river bank is artificially constructed for river training, and there are no coastal mangroves or tidal mudflats. The regularly flooded marshes represent the ponds that support reedbeds in the Nature Reserve, and under a 2 mm/yr accretion rate, over 50% of these marshes will be lost, indicating that some of the ponds that are not connected to the estuary via tidal creeks will also become inundated during low tide under a 1.5 m SLR (Table 2). This likely reflects erosion and subsidence of the tidal flats, allowing sea water intrusion into the nature reserve.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. SLAMM outputs derived from this analysis to show changes to habitats at low tide under 1.5 m sea level rise and 15mm, 8mm, 4mm and 2mm/year accretion rates by 2100.

Landuse basemaps used in the analysis are from WWF Hong Kong https://globil-panda.opendata.arcgis.com/datasets/mai-po-wetland-habitats.

https://doi.org/10.1371/journal.pone.0239945.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Changes to habitats under the 1.5 and 2 m sea level rise scenarios and different accretion rates by 2100 as projected by the SLAMM.

https://doi.org/10.1371/journal.pone.0239945.t002

Under a more extreme 2 m SLR scenario, the tidal flats and coastal mangroves will still remain relatively unaffected during low tide at accretion rates of 15 mm/yr, but there will be a loss of about 77 and 57%, respectively, of tidal flats and mangroves with an 8 mm/yr accretion rate (Table 2). The regularly flooded marshes in the nature reserve will also lose 56% of area with an 8 mm/yr accretion rate, and the inland open water marshes will lose about 18% of area even with a 15 mm/yr accretion rate.

Under the high tide scenarios, with 1.5 m SLR and 15 mm accretion levels about 76% of the tidal flats will become inundated, while the mangroves will remain unaffected (Table 2, Fig 3). There will be a loss of between 19 to 22% of inland open water marshes during high tide, a marginal increase from the low tide scenario (Table 2). However, over half of the regularly flooded marshes inside the nature reserve will also be inundated even with an 8 mm/yr accretion rate and less (Table 2). Under a 2 m SLR scenario, during high tide over 77% of the tidal flats and about half of the mangroves and marshes will be inundated even with a 15 mm accretion rate (Table 2). If the accretion decreases to 4 mm/yr or less, most of the marshes, mangroves, and tidal flats will be underwater.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. SLAMM outputs derived from this analysis to show changes to habitats at high tide under 1.5 and 2 m sea level rise and 15mm and 2mm/year accretion rates by 2100.

Landuse basemaps used in the analysis are from WWF Hong Kong https://globil-panda.opendata.arcgis.com/datasets/mai-po-wetland-habitats.

https://doi.org/10.1371/journal.pone.0239945.g003

Discussion

Long distance animal migrations are some of Nature’s more wondrous phenomenon. But across the world these migrations are threatened by habitat loss, barriers that constrain or prevent movement, hunting [38], and now, climate change [11]. The EAAF is one of nine major long-distance corridors for migratory waterbirds that covers a swath over East, Southeast and eastern parts of South Asia to include 22 countries from Arctic Russia and USA (Alaska) to Australia and New Zealand [39]. The wetlands that lie scattered through this vast landscape, especially along the coastal areas, represent steppingstones where the waterbirds, which comprise a large group of the migrants, can rest, refuel, and continue on their journey, or provide winter habitats [40].

But the ecological functionality of the flyways that have evolved over millennia are now threatened [6, 8, 9, 41, 42]. Asia’s coastal areas are among the most densely populated areas, and urban and industrial infrastructure is rapidly expanding at the expense of the natural ecosystems disregarding the importance of the ecosystem services these wetlands provide and the longer term economic and social consequences of losing them [5, 8, 43–46]. Over 80% of East and Southeast Asia’s wetlands are now classified as threatened and nearly half of all tidal flats have already been lost [4]. Thus, many of the migratory waterbirds that depend on the wetlands along the EAAF are now facing population declines as important and strategic stopover and winter sites continue to disappear [1, 4, 8, 35]. While over 5 million individuals from 58 species of shorebirds are known to use the EAAF, it now has the highest number of threatened species and declining populations [47]. Analyses have shown that less than 10% of the wetlands that support waterbirds are protected, and many important yet unprotected areas are vulnerable to anthropogenic threats [3, 34, 48]. Thus, given the rapid loss of wetlands, inadequate protection and conservation management and consequent erosion of ecological connectivity and functionality of the flyway, it is important to secure the remaining wetlands to ensure that the migrations persist [35].

In addition to protection to prevent anthropogenic conversion, the wetlands and the flyway also require a ‘climate-smart’ conservation paradigm [49–51]. As transitional ecosystems between terrestrial and marine systems, coastal wetlands are vulnerable to SLR and the severe storms and associated strong wave surges expected from climate change on one side, and continued land conversion in the terrestrial realm [52, 53]. With SLR, coastal wetlands could migrate further inland [48, 54]. But if the adjacent terrestrial areas are converted to anthropogenic land uses this migration will be constrained and the wetlands will be lost due to coastal squeeze [48, 55]. As strategic wetlands along the flyway are lost to development and climate change and consequent environmental impacts, the functionality of the flyway will become degraded, and the migration phenomenon could eventually cease.

Acknowledgments

We thank the Peter Cornthwaite, CEO WWF Hong Kong and David Olson, Conservation Director for encouraging this study, Chi-Yeung Choi and a second anonymous reviewer for constructive comments. We also thank the Mai Po management team and management committee for their support.