Freshwater ecosystems (or inland wetlands) include marshes, swamps, peatlands, wetland forests, rivers, lakes, ponds and headwaters. They provide a range of provisioning, regulatory and supporting ecosystem services, including water and food supply, fodder and building materials, carbon and nutrient sequestration, unique habitats for endangered species (including migratory birds), flood- and drought-buffering capacity, ecotourism and cultural services (WWAP 2018). Although freshwater ecosystems only cover 0.8 per cent of Earth’s surface, they support approximately 10 per cent of all known species (World Wide Fund For Nature [WWF] 2016), and are among the world’s most biodiverse habitats. They are the ecosystems most affected by changing land use, particularly increasing urbanization and agricultural expansion.
Between 69 per cent and 75 per cent of wetlands worldwide are estimated to have been lost since 1900 due to rapid population growth, urbanization and agricultural expansion (Davidson 2014). The extent of the loss since 1970 differs notably across regions, the slowest loss rate being apparent in Oceania and North America. Levelling off of the loss rate in North America is due partly to the current national policy of “no net loss of wetlands” in the United States of America (United States Fish and Wildlife Service 1994). Although constructed wetlands can compensate to some degree for some natural wetland removal, they cannot typically provide the same level of ecosystem functioning, resilience and biodiversity, emphasizing the need for natural wetland protection and conservation (see Section 9.4).
Ecosystem services for all wetland types have been valued financially across a very wide range from US$300-US$887,828 per hectare per year, with a median value of US$12,163 (de Groot et al. 2012). More specific assessments are needed. A recent valuation of swamp and floodplain ecosystem services attributed an average annual global value of US$25,000 per hectare per year, excluding the value of the land itself (Costanza et al. 2014). The estimated annual loss to the global economy from diminishing swamp and floodplain areas from 165 to 60 million ha between 1997-2011 is US$2.7 trillion (Costanza et al. 2014).
Although covering only 3 per cent of the planet’s land surface, peatlands have a high carbon-sequestration value, hence they contain more carbon than all global forest biomass combined (Joosten 2015). The world’s largest tropical peatland (Cuvette Centrale) covers an area of 145,500 km2 in the Congo River basin, containing an estimated 30 gigatons of carbon accumulated over the past 11,000 years (Dargie et al. 2017). Draining peatland areas for agriculture (e.g. the large Indonesian and Malaysian palm oil plantations) breaks down the peat, rapidly emitting carbon as CO2 and methane. About 15 per cent of peatlands worldwide have been drained in the last 40 years, contributing approximately 5 per cent of global carbon emissions (Joosten 2015).
As the drained peat decomposes quickly, it dries out, shrinks and subsides. Tropical coastal peatlands are subsiding by an average 5-7 cm/year, and thus become vulnerable to salinization during storm surges. During hot, dry periods, the fire hazard in peatlands is high (Jayachandran 2009), an example being extensive peat fires in Indonesia exacerbating brown-haze pollution of the whole Asian region in the summer of 2015 (Carmenta, Zabala and Phelps 2015).
The permafrost in boreal peatlands in and around the Arctic Circle is thawing and draining due to climate change, with effects on local and global carbon fluxes (Joosten 2015; Couture et al. 2018). Apart from additional emitted carbon, the permafrost thawing is damaging infrastructure and housing, affecting Arctic people’s quality of life. For both tropical and boreal peatlands, the straightforward technical solution to addressing carbon emissions from drained peatlands is to rewet the peatland, bringing the water table back to the soil surface, as is currently being done at a large scale in Indonesia, Canada, Sweden and Switzerland (Zerbe et al. 2013).