Colonization of UK coastal realignment sites by mosquitoes: implications for design, management, and public health



Coastal realignment is now widely instituted in the UK as part of local flood risk management plans to compensate for the loss of European protected habitat and to mitigate the effects of sea-level rise and coastal squeeze. Coastal aquatic habitats have long been known to provide suitable habitats for brackish-water mosquitoes and historically, coastal marshes were considered to support anopheline mosquito populations that were responsible for local malaria transmission. This study surveyed the eight largest managed realignment (MRA) sites in England (Essex and the Humber) for mosquito habitats. The apparent absence of anopheline mosquitoes exploiting aquatic habitats at all of these sites suggests that the risk of malaria associated with MRA sites is currently negligible. However, three of the eight sites supported populations of two nuisance and potential arboviral vector species, Aedes detritus and Aedes caspius. The aquatic habitats that supported mosquitoes resulted from a) specific design aspects of the new sea wall (ballast to mitigate wave action and constructed saline borrow ditches) that could be designed out or managed or b) isolated pools created through silt accretion or expansion of flooded zones to neighbouring pasture. The public health risks and recommendations for management are discussed in this report. This report highlights the need for pro-active public health impact assessments prior to MRA development in consultation with the Health Protection Agency, as well as the need for a case-by-case approach to design and management to mitigate mosquito or mosquito-borne disease issues now and in the future.


The rise in sea levels around the coasts of the UK is being driven in two ways. Recent studies show that the ice sheets in Greenland and Antarctica are melting faster than the snow is replacing the mass. In their fourth report (Solomon et al. 2007), the Intergovernmental Panel on Climate Change predicted a global sea level rise increase of 18–38 cm by 2100 in the most optimistic scenario and 26–59 cm in the most pessimistic. In the UK, the effect of this sea level rise is further exacerbated by the effect of glacial isostatic adjustment (or post-glacial rebound/depression). This phenomenon is related to post-glacial activity whereby northern Britain undergoes an uplift following deglaciation and southern Britain sinks, particularly in southern and eastern England. Predictions on sea level rise have been made by the UK Climate projections (UKCP09), supported by the UK Department for Environment, Food and Rural Affairs. Incorporating both these factors, the central estimates for London of sea level rise with respect to 1990 levels predict a 5.3–7.3 cm rise by 2010, 8.2–11.5 cm rise by 2020, 18.4–25.8 cm rise by 2050, and 30.5–43.3 cm rise by 2080 (Murphy et al. 2009).

This presents challenges to coastal stability, particularly in the face of storm surges (Murphy et al. 2009). This has led to the implementation of managed realignment (MRA) schemes along the coast which essentially involve replacing artificial hard coastal defenses with natural soft defenses like coastal habitats (Leggett et al. 2004). MRA is the deliberate process of altering flood defenses to allow flooding of a presently defended area. It is now considered to offer long-term sustainable management of coasts and estuaries for a variety of stakeholders. It can reduce the pressures of coastal squeeze and offer potential new habitat creation and re-creation opportunities (Leggett et al. 2004).

An MRA scheme may take many forms, including a) retreating to higher ground, i.e., where a line of defense is breached/removed, allowing inundation up to the higher ground, producing a new intertidal area, b) constructing a set-back line of defense, thus protecting property landward of the defense, or land at lower altitude to the defense, c) shortening the overall defense length to be maintained, thus reducing costs or d) inundation through sections of defense whereby the breach allows inundation of land behind it through a defined gap, or through pipes with tidal flaps to allow an intertidal area to develop behind the defense.

The flood management and environmental benefits of MRA are manifold (Leggett et al. 2004). These include a) reducing flood risk or potential for reduction in whole life costs of flood defense, b) ensuring long-term sustainability of defenses by increasing natural flood and storm-buffering capability, c) reducing the costs, particularly where existing defenses are not economically viable, d) assisting with managing the effects of sea level rise by reducing height of sea/estuary levels, and e) mitigating the effects of previous reclamations and of climate change (i.e., sea level rise), and thus offsetting the impact of coastal squeeze. Most significantly, they lead to creation of intertidal habitat, usually mudflat and salt marsh with the associated benefits for wildlife and protected species, as well as improving recreational value.

MRA schemes in the UK addressing coastal squeeze are largely initiated by the Environment Agency (EA) as part of their flood defense remit. However, they may also arise as compensation sites for direct loss usually through the Environmental Impact Assessment (EIA) process and/or as part of an Appropriate Assessment under EU legislation. For the latter schemes, these are initiated by the prospective developer, in this case the competent authority, Associated British Ports (Mander et al. 2007, Hemingway et al. 2008). In addition to the UK, MRA is now being widely used at coastal sites in France, Belgium, the Netherlands, Germany, and Denmark. Their impact on mosquitoes is therefore a concern for a number of European countries.

With respect to mosquitoes however, salt marsh habitat and coastal marshes (both favoring brackish waters) have been, and are currently, responsible for providing habitats for nuisance biting mosquitoes and, albeit historically, disease vector species. Many of the sites in coastal Essex and Kent where malaria was endemic in the 19th century were drained, and this was considered to be a major contributory factor to the decline of malaria in England (Dobson 1997). The most recent outbreaks of indigenous malaria in the UK in the 20th century were associated with coastal populations of Anopheles atroparvus (Lindsay and Willis 2006). In certain parts of the UK, the salt marsh mosquito Aedes detritus is a significant biting pest (Medlock et al. 2012) and an economic burden to local councils. Increased temperatures in southern England as a consequence of climate change, which could favor pathogen development, coupled with the increase in imported cases of mosquito-borne disease and the emergence of exotic arboviruses in Europe suggest that we should not be complacent when considering the creation of new brackish and coastal habitat in England with respect to public health (Medlock and Leach 2012).

Aspects that exacerbate the potential nuisance caused by salt marsh mosquitoes are related to a) the high marsh where pools of water in mud flats or salt marsh vegetation are left by the highest tides, or alternatively are filled by rainfall/runoff or not flushed by daily tide movements, and b) the low marsh that is not well drained and where mosquitoes exploit impounded stagnant pools that are retained, usually due to siltation/blockage of tidal channels and hence not flushed (Russell 2009). Management strategies include the elimination of the potential aquatic habitat (by draining or filling), modification (with water management), and treatment with a control agent to kill the mosquito larvae. Elimination is usually not possible, but modification with strategies such as ‘Open Marsh Water Management’ (OMWM) or the use of shallow ditches (runnels) has been reported to be acceptable, practical, and effective (Russell 2009).

OMWM was developed to control mosquitoes by introducing their natural predators to areas of salt marsh. With a system of pools connected by radial ditches (i.e., runnels), fish feed on mosquitoes during high tide, then retreat to sumps or reservoirs at low tide. OMWM has been found to be an effective long term method of controlling mosquito populations in salt marshes without using sprays (SWS 2009). This strategy promotes/restores ‘full tidal flushing’ by advocating the renovation of tidal channels and maintaining them in a condition which allows a) full tidal exchange and precludes the formation of impounded pools and b) ‘natural dewatering,’ whereby salt marsh pools that hold water after highest tides and rainfall are connected for tidal influence using various sized channels and with persisting ponds to support predatory fish (Russell 2009).

Since 1995, there have been at least twenty coastal realignment projects on the east coast of Britain between and including the Thames and Humber estuaries, the largest of which covers >350 hectares (Hemingway et al. 2008). Some of these sites are in areas where malaria was previously common. Such newly created habitats have the added benefit of providing additional breeding grounds for birds, and the combination of wild birds and mosquitoes is now widely accepted as increasing the transmission risk of bird-associated viruses such as West Nile virus (Medlock et al. 2005).

Despite significant biodiversity monitoring of these MRA sites post-sea incursion (Hemingway et al. 2008), there has been no assessment of the colonization by mosquitoes or the associated potential disease risks. Some sites are in remote parts fringing large estuaries, with little or no neighboring human settlement. However, new sites are being designed and planned for the south coast of England where human populations are significantly higher. A baseline understanding of the possible nuisance biting and public health risks associated with such coastal habitat creation is now required (Medlock and Vaux 2011), particularly if the design of new sea defenses are found to inadvertently create unwelcome mosquito habitats.

This study surveys the largest eight MRA initiatives in Essex and the Humber to ascertain whether mosquitoes are able to colonize them. Furthermore, by comparing recent schemes with those built in the 1990s, it may be possible to semi-qualitatively assess the rate of colonization of mosquitoes. Consideration will also be given to the impact of new sea wall design on creating new aquatic mosquito habitats and an assessment of the potential public health risks, with suggestions made on design mitigation strategies to limit potential nuisance and disease risks.


Field sites on the Humber

Field visits were made to eight of the largest coastal realignment sites in the UK, four on the Humber Estuary in Yorkshire and North Lincolnshire and four in Essex (Table 1, Figure 1). On the banks of the Humber, Paull Holme Strays, Welwick, and Chowder Ness were all designed with the construction of a new sea wall set back from the estuary with one or more breaches in the old wall. All three sites receive highly saline water input and are flooded on each high tide and almost entirely drained at low tide. The dominant habitat created at Welwick and Chowder Ness is mudflat, with areas of salt marsh developing by the new sea wall; particularly at Welwick. The sea wall at both of these two sites is constructed with a subsidiary bank of ballast on the seaward side, presumably to mitigate the effects of the high water and wave action.

Table 1.  Overview of managed realignment sites surveyed.
SiteLocation (Coordinates)SizeDate of constructionHabitats created and driver
  1. * EA (Environment Agency), ABP (Associated British Ports), RSPB (Royal Society for Protection of Birds), EWT (Essex Wildlife Trust), NRA (National Rivers Authority).

Paull Holme StraysEast Yorkshire (53.7N, 0.2W)80 ha2003 (EA)45ha mudflat, 35ha saltmarsh; address coastal squeeze and improve flood risk management
Chowder NessLincolnshire (53.7N, 0.5W)15 ha2006 (ABP)Mudflat; compensation for impact of creating new harbour
WelwickEast Yorkshire (53.6N, 0.03E)54 ha2006 (ABP)18–38ha mudflat, 12–28 ha saltmarsh; compensation for impact of creating new harbour
AlkboroughLincolnshire (53.6N, 0.6W)440 ha2006 (EA)375ha of estuarine habitat; flood storage and flood risk management
Wallasea IslandEssex (51.6N, 0.8E)125 ha2006 (RSPB)90ha mudflats, 23ha saltmarsh, 6ha freshwater marsh; replace lost habitats for conservation
Abbotts HallEssex (51.8N, 0.8E)120 ha1999 (EWT)120ha of estuarine habitat; replace habitats lost for conservation, ageing sea defences
OrplandsEssex (51.7N, 0.9E)38 ha1995 (NRA)New habitats for wildlife, mitigate storm surges
TollesburyEssex (51.7N, 0.8E)21 ha1995 (NRA)Mainly mudflat habitat
Figure 1.

Location of managed realignment sites (black dots) surveyed a) on the Humber estuary and b) on the Essex coast. Main towns are represented by grey dots, with size representing their human population c) locations within Great Britain.

Chowder Ness forms part of the south Humber Bank series of mainly freshwater wetlands close to Barton-Upon-Humber. Welwick lies adjacent to arable farmland and the silted up entrance to Patrington Haven, an area of coastal flooded grassland. Paull Holme Strays is a more complex site with a series of islands (outstrays), areas of mudflat, and developing salt marsh. It is located between the city of Hull and arable farmland, with limited neighboring coastal habitat.

The fourth site on the Humber, Alkborough, is located at the head of the estuary at the confluence of the rivers Ouse and Trent, where the input is mainly freshwater. Water enters the site through a single narrow breach, and the inundation area is not restrained by a sea wall, although there is a flood bank some way from the inundation zone. The neighboring habitat in the east is mainly cattle pasture, and to the south there is a new area of recently created freshwater habitat (i.e., reed-beds, ponds, and ditches) as well as inundated ash woodland.

Two of the sites (Welwick, Chowder Ness) were constructed as part of a compensation package during the development of a new roll-on roll-off terminal at Immingham Harbour. The other sites (Paull Holme Strays, Alkborough) were developed as part of the Humber flood risk management plan and to mitigate coastal squeeze. Alkborough was additionally constructed to reduce water levels in the upper estuary during extreme events, being able to reduce high tide levels by 15 cm, in order to minimize flooding in the city of Hull.

Field sites in Essex

In Essex, two of the sites (Orplands, Tollesbury) were constructed in 1995 and were pioneering projects at the time. They were initiated to mitigate the effects of storm surges in the Blackwater Estuary and its tributaries and provide additional habitat for wildlife and recreation. Tollesbury is an extension of Tollesbury Fleet and is surrounded by extensive salt marsh and mudflat habitats. The site is almost entirely inundated by the tide. Orplands lies on the main Blackwater estuary and is largely surrounded by arable farmland with some extant salt marsh in the vicinity at St. Lawrence creek. The old sea wall at Orplands has been breached in several places with salt marsh habitat developing behind, constrained by the new sea wall.

The other sites (Abbotts Hall, Wallasea Island) were largely initiated by conservation organizations with the aim of providing habitats for birds. However, they also had additional benefits in assisting flood risk management and replacing aging sea defenses. Abbotts Hall is situated on the north bank of the Salcott Channel, a tributary of the Blackwater estuary. The old sea wall has been breached in five locations, with mainly salt marsh habitat developing behind. The MRA site at Wallasea lies on the south bank of the River Crouch estuary and forms only a small part of a much greater project to return the whole of the island to coastal habitat. In the MRA site, mudflat is dominant, with some salt marsh. However, pipes through the new sea wall allow inundation by saline water of a large 3 km borrow-ditch created by sea wall construction, lying immediately parallel with the sea wall. The grading of this ditch varies so that some areas remain permanently wet and others are flooded at each high tide or only at spring high tides.

Survey strategy

Two field visits were made to each of the eight sites during March/April and August, 2011 to coincide with the main activity period of the immature stages of coastal mosquitoes (principally An. atroparvus, Ae. detritus, Ae. caspius, and Cx. modestus). All aquatic habitats at each site were visited on both occasions and using standard mosquito dippers (5 × 200 ml dips per location), the occurrence of mosquito larvae was recorded. Aquatic habitats were defined as new aquatic habitat within the MRA area that is either within the inundation zone or at the margin of the inundation zone, as new aquatic habitat adjacent to the inundation zone within the MRA boundary (always freshwater habitats), or as extant aquatic habitat within the boundary of the MRA. All samples were returned to the laboratory for identification of mosquito species in accordance with Snow (1990) and Schaffner et al. (2001), and abundance was recorded by species and stage. Locations of aquatic habitats surveyed were recorded using a handheld GPS datalogger with further information recorded on salinity, temperature, and vegetation coverage. Each survey site was classified on the basis of its geographic location (Essex or Humber), whether the site had a subsidiary bank in front of the sea wall (lip or no lip), the salinity of the water (fresh or salt), and the impact of the tide (regular flood zone or marginal isolated pools). These constituted the explanatory variable associations between mosquito presence (the response variable) and were tested using chi-squared, Fisher's exact and regression tests using Minitab 15 (State College, PA, U.S.A). In addition, and in order to make comparisons of new coastal habitats with extant coastal habitats, an additional extant salt marsh site at Seymer's Marsh on Brownsea Island, Dorset (50.7N, 2.0W) was surveyed. This site was initially studied by Service (1968) and revisited by the authors in February, 2011.


Almost 400 aquatic habitats were sampled across the eight MRA sites, with the number of samples and number of positive mosquito samples presented in Table 2. The majority of mosquitoes were associated with the tidal-influenced aquatic habitat in the MRA (i.e., not including extant habitat or newly created permanent freshwater habitats), with the highest numbers at Alkborough and Welwick, both on the Humber. Aedes caspius dominated at Alkborough and Ae. detritus at Welwick. Aedes detritus was also found in the saline borrow ditch at Wallasea Island (Table 3). Additionally, Ae. detritus was found in extant habitat neighboring the MRA site at Abbotts Hall, with Culex pipiens and Ae. caspius in extant habitat neighboring Alkborough. Anopheles maculipennis s.l. and An. claviger were found in newly created permanent freshwater ditches and ponds adjacent to the inundated zones at Alkborough. Further details of the specific positive locations at each site are discussed briefly below.

Table 2.  Analysis of mosquito sampling at managed realignment sites. ‘+ve sites’ refers to survey sites with mosquito immatures present; ‘within MRA’ refers to all sites that have been subject to managed realignment and therefore impacted by tidal waters; ‘adjacent MRA’ refers to additional freshwater habitat created within the construction site but not affected by tidal waters (i.e., beyond the new sea wall).
LocationCountyDatesSalinity range (mean±SE)Total no. aquatic sites surveyedTotal sites +ve for mosquitoesn sites within MRA+ve sites within MRAn sites adjacent MRA+ve sites adjacent MRA
Abbott's HallEssex15 Apr, 16 Aug1.85±0.27552280232
AlkboroughLincolnshire17 Mar, 9 Aug0.016±0.00513211797534
Chowder NessLincolnshire18 Mar 8 Aug0.13±0.03303000
OrplandsEssex4 Apr, 16 Aug2.44±0.233033000
Paull Holme StraysYorkshire16 Mar, 5 Aug1.10±0.0658056020
TollesburyEssex6 Apr, 16 Aug1.87±0.61606000
WallaseaEssex4 Apr, 16 Aug2.00±0.227123140
Welwick AYorkshire16 Mar, 5 Aug1.21±0.115114501410
Welwick BYorkshire18 Mar, 5 Aug1.85±0.08227227
Table 3.  Abundance of mosquitoes by species at managed realignment sites.
LocationInside/Adjacent MRA – site specificSpeciesStage (I–III/IV/P)
Abbott's HallExtant saline pool adjacent to MRA site Ae. detritus 22/0/0
AlkboroughNewly constructed freshwater ditches/ponds adjacent to MRA An. claviger
An. maculipennis s.l.
AlkboroughFlooded margin of ditch behind seawall, adjacent to MRA Cx. pipiens s.l.0/0/1
AlkboroughFlooded cattle pasture, beyond mudflat at margin of MRA Ae. caspius
Cx. pipiens s.l.
AlkboroughReed dominated pool next to mudflat at margin of MRA Ae. caspius
Cx. pipiens s.l.
WallaseaIntermediate bank of saline borrow ditch, behind sea wall, margin of MRA Ae. detritus 2/2/4
Welwick AAmong vegetation and rocks in ballast lip below sea at margin of MRA Ae. detritus 144/36/7
Welwick BIn extant saline pools with emergent reed, at margin of high tide, adjacent to MRA Ae. detritus 41/0/0


The MRA site at Welwick (hereafter Welwick A) is largely mudflat and almost entirely inundated by saline waters at high tide (visits were made at high and low tide) and consequently was largely unsuitable for immature mosquitoes. The extant salt marsh habitat on the unbreached section of the old coastline is regularly flushed with no evidence of pooling to support mosquito aquatic habitats. However, adjacent to the new sea wall, brackish water collects in front of, among, and behind a subsidiary bank of ballast. All mosquitoes found at this site were located here and all were Ae. detritus, found in substantial numbers both in March and August. Specific aquatic habitats included water among un-vegetated ballast, water in a linear depression between the ballast and the sea wall, and among flooded halophytic vegetation on the seaward side of the ballast. All these pools appear to be the result of pooling after spring tides or from spray. These were isolated from the estuarine waters even at high tide (Figure 2a).

Figure 2.

Typical aquatic habitats for mosquitoes at coastal realignment sites. a). Welwick: pools above high tide adjacent to subsidiary ballast bank supporting Ae. detritus. b). Alkborough: pools in cattle pasture, left isolated by spring tides supporting Ae. caspius. c). Alkborough: pools among developing reed, at margin of mudflat supporting Ae. caspius. d). Wallasea: vegetated terrace adjacent to main saline ditch in excavated borrow-ditch supporting Ae. detritus.

Additional surveys were conducted in an adjacent area of coastal aquatic habitat (hereafter Welwick B), reported to be the silted up entrance of Patrington Haven. This area is bound by a sea wall, and includes an area of 500 m wide salt marsh and saline-flooded coastal grassland. Aedes detritus was also found at this site in March but was only associated with areas of flooded grassland or vegetated pools at the spring high tide mark next to the sea wall.


Tidal waters enter the Alkborough site through a single breach and much of the inundated zone is developing mudflat. It was noticeable that accretion of the mudflat was much higher close to the entrance of the breach channel, with water collecting behind this as more permanent shallow water at locations furthest from the breach. Salinity levels were extremely low with input largely of freshwater from the rivers Trent and Ouse. The fringes of the inundated zone are not bound by a sea wall, so there is significantly more emergent vegetated habitat at this site. Despite considerable sampling effort, immature mosquitoes were only found in two locations, each as a number of distinct aquatic pools at each location. Water from spring high tides had flooded cattle pasture on the eastern margin of the site, with a number of isolated linear open flooded sites on the grassland, which were nutrient rich with cattle dung and significantly puddled by animals. Larvae of Ae. caspius were found here in very high numbers in the drying pools (Figure 2b). On the western fringe, although mudflat water levels at low tide were very low (<3 cm), several deep pools of isolated water are retained among the reed at low tide. It is possible that reed growth here has facilitated higher accretion on the margins of the high water mark, thus isolating collections of water left by the highest tide. These aquatic habitats include narrow runnels within reed or open pools fringed with emergent reed. They contained significant numbers of immature Ae. caspius (Figure 2c).

Elsewhere in neighboring extant aquatic habitats there were very low numbers of Ae. caspius in flooded grassland in March. Alkborough additionally has a number of newly created freshwater habitats, and low numbers of An. claviger, An. maculipennis s.l., and Cx. pipiens s.l. were found in ditches and ponds. There were no immature larvae associated with the flooded woodland.

Wallasea Island

No immature mosquitoes were found in the MRA area between the new sea wall and the river. The new habitat is predominantly mudflat and the marginal vegetation under the sea wall is regularly flushed by the tide, both inimical for mosquitoes. All saline pools in the extant habitat were also negative for mosquitoes. The spoil used to create the new sea wall was procured from a linear borrow ditch which is wetted and connected to the main river through a sluice tunnel. This ditch does dry out in places and during high water levels, floods a more elevated vegetated bank within the main culverted ditch. Aedes detritus larvae were found in drying pools among vegetation on this bank (Figure 2d).

Abbotts Hall

No immature mosquitoes were found throughout the MRA site. There is a large amount of flooded grassland, but all of this appears to be flushed by the tide, with no isolated pools. First instar Ae. detritus were found in highly saline pools by the jetty. It is questionable whether this population would develop to imagoes as there was no emergent or floating vegetation or other fauna in these pools.

Chowder Ness

Much of this site is mudflat, with some marginal vegetation below the new sea wall and among the subsidiary bank of ballast and flotsam washed in by the tide. In contrast to Welwick these sites were largely dry and evidenced by the flotsam high on the sea wall; the tide regularly washes above this ballast bank. No mosquitoes were found on either visit.


Water is retained within the MRA site even at low tide. Water collects behind the sea wall at low tide, and the well-developed salt marsh has numerous runnels with exposed mud and flooded vegetation. As none of the pools are isolated, it is likely that they all receive regular tidal flushing making them inimical for immature development. No mosquitoes were found on either visit.


This site is almost all mudflat with only low level vegetation developing on mud by the sea wall. Very few aquatic pools were found at low tide, and on the return visit all were inundated at high tide. This site is currently unsuitable for immature mosquito development.

Paull Holme Strays

The western part of the site is predominantly mudflat. With an absence of a subsidiary bank of ballast below the seawall, no isolated pools appear to collect that could support mosquitoes. The eastern part is more complex and more vegetated. There are a number of wetted areas and pools in the grassland, but all were devoid of immature mosquitoes. Flooded grassland areas on the outstray were also negative for mosquitoes. It is possible that some of these aquatic habitats could support mosquitoes, but so far they have been unable to colonize, possibly due to low dispersal rates.

Summary of results and statistics

The occurrence of mosquitoes within the MRA area was significantly associated with the presence of a subsidiary bank on the sea wall (bank: mosquitoes present at 14/53 aquatic sites; no bank: 8/225; P<0.001), isolated pools distinct from the regular inundation zone (flood zone pools: 0/210; isolated pools: 22/68; P<0.001), and sites on the Humber (Humber: 21/188; Essex 1/90; P=0.0015). In a stepwise regression analysis, the combination of a) no regular tidal impact and b) the presence of a bank in a model was significantly associated with the occurrence of mosquitoes (R2(adj)= 31.23; P<0.001).

Although the salinity of the water was not a contributory factor to mosquito presence (freshwater pools: mosquitoes present at 7/79 aquatic sites; brackish pools: 15/199: P>0.05), it did determine the species of colonizing mosquito. The occurrence of Ae. detritus was significantly associated with brackish sites (freshwater pools: 0/79; brackish pools: 15/99; P<0.001) and the occurrence of a subsidiary bank (bank: 14/53; no bank: 1/225; P<0.001). Using a stepwise regression analysis combining these two variables gave a model with R2(adj)= 30.83 (P<0.001).

Aedes caspius was significantly associated with freshwater sites (freshwater: 7/79; brackish: 0/199; P<0.001). With respect to anopheline species and the risk of malaria, a further analysis was conducted on whether MRA sites significantly support Anopheles mosquitoes. Significantly more anophelines were found in the constructed permanent freshwater habitats outside the main MRA inundation zones (P=0.05).

Results from extant salt marsh habitat

The saline lagoon at Seymer's marsh is regularly flushed by the tide, and the degree of marginal vegetation varies depending upon the tidal velocity and the distance from the main incoming tide. It is a complex site with many narrow channels. Immature Ae. detritus were found at 25% (25/101) of the survey points, and its occurrence was compared with six explanatory variables: salinity, temperature, depth, size of water body, vegetated vs mud substrate, and occurrence of rush/reed (Table 4). A stepwise regression analysis showed a significant association between mosquito presence and the occurrence of smaller water bodies, and a vegetated substrate (R2(adj)= 16.1, P=0.017)

Table 4.  Survey data in saline lagoon at Brownsea Island, Dorset.
Predictors+ve site for mosquitoes-ve site for mosquitoes
n = 1012576
Salinity (%)1.94±0.142.23±0.76
Temperature (°C)8.91±0.159.02±0.06
Depth (cm)8.92±1.6511.01±0.94
Size of Waterbody24.14±9.34105.7±11.9
Vegetated substrate0.44±0.100.19±0.04
Presence of rush/reed0.92±0.0550.63±0.56


Historically, the occurrence of indigenous malaria in the UK was associated with coastal marshes, usually grazing marshes. One of the reasons for the decline in malaria in the UK was that many of these coastal wetlands were drained for agriculture (Lindsay and Willis 2006). Concern has been raised that creating new coastal wetlands, such as those created through managed realignment, would heighten the risk of malaria transmission. This study aimed to answer this question, as well as to understand the occurrence of other mosquito species that might be implicated in the transmission of other mosquito-borne pathogens, for example West Nile virus (Medlock et al. 2005, Medlock et al. 2007). However, it may be possible to manage such mosquito populations through hydrological and vegetation management (Medlock and Vaux 2011), or indeed ensure that the design of such schemes does not inadvertently create aquatic habitats for nuisance and disease vector mosquito species.

In conclusion, based on these surveys, there is no current evidence to suggest that MRA sites heighten the risk of malaria transmission. There was no evidence of any Anopheles mosquitoes (putative malaria vectors) at any of the sites where tidal inundation occurred. It is possible that at other sites, where coastal grazing marshes are created, the situation might be different. Anophelines were found in newly constructed freshwater ditches and ponds, but this is no more than one would expect from any other permanent pond or ditch habitat (Medlock and Vaux 2011). As these anopheline populations were not in the managed realigned area, no further molecular diagnostics were conducted.

All mosquitoes found with the MRA sites were associated with isolated pools at the extremes of the flood zone. For Welwick, all aquatic habitats for Ae. detritus were associated with the additional subsidiary bank of ballast in front of the new sea wall. Unless this is regularly flushed by the tide, as appears to happen at Chowder Ness, this additional bank will support significant numbers of mosquitoes, as occurs at Welwick. The need for this additional bank, and the frequency with which it is flushed by the tide (thus making it inimical for mosquitoes) should be a consideration in the design and tidal-modelling phase. Aedes detritus is an important nuisance species, and elsewhere in coastal sites (Dee estuary, Sandwich Haven) is a subject of ongoing mosquito control initiatives (Medlock et al. 2012). Owing to the confined habitats of this mosquito, it would be possible to institute regular larvicidal control of these habitats if they were deemed to be causing a nuisance.

The situation at Alkborough is unique for an MRA site in the UK, in that the incursion is largely freshwater. The water enters by a narrow breach, flooding farmland not immediately proximal to the estuary. This presents several opportunities for mosquitoes. Firstly, owing to the freshwater input, this allows fresh floodwater mosquitoes to colonize, in this case, Ae. caspius. Secondly, as the input of water is through a narrow breach, and owing to the level of accretion close to the breach, not all tidal waters are drained at low tide. This is not a problem per se, as areas that are regularly flushed will not support mosquitoes and particularly large areas of water with consistent surface wind movement are also unsuitable. However, water that regularly flushes a habitat at low velocity assists with marginal silt build-up, particularly in areas where silt is trapped by marginal reed growth. This can create isolated pools that support mosquitoes. Although they may occasionally be flushed by spring tides, their location offers enough protection by vegetation at low water velocities to support mosquitoes. If so decided, these aquatic pools could be drained by digging a narrow trench through the reed towards the mudflat. At Alkborough, this is probably not currently required. Finally, a greater area of Alkborough is flooded than was originally predicted by the modelling. This has led to flooded woodland that supported no mosquitoes, but also a series of flooded hedges and cattle pastures. A flooded cattle pasture (provided that the water is quickly isolated from the main tidal flushing) provides the perfect aquatic habitat for Ae. caspius, with cattle acting as an ideal blood source. Again, such habitats could be controlled through either draining or larvicidal control.

The occurrence of mosquito larvae at Wallasea Island illustrates a different issue. While no mosquitoes were found on the seaward side of the sea wall, the flooding of a posterior ditch with tidal saline waters provides a low water velocity habitat for mosquitoes. It is possible that such ditches could support anopheline mosquitoes, as they mimic their favored coastal grazing marsh habitat, however none were found. The construction of an elevated bank within the ditch, that only occasionally floods, is perhaps ideal from a mosquito habitat perspective, and it is very likely that this habitat explains the nuisance biting reported here. Moderating the flow of tidal waters through the sluice during spring and summer to a level below this elevated bank might negate such habitats for mosquitoes and consequently remove the nuisance issue.

The absence of mosquitoes at the other sites is largely explained by the almost complete tidal flushing of all aquatic habitats. However, the absence of mosquito larvae in isolated pools at Paull Holme Strays is surprising and may be accounted for by the absence of inward-dispersing mosquitoes. The site is surrounded by the city of Hull and arable land, and perhaps in time, mosquitoes could colonize this site and become established.

As previously mentioned, regular tidal flooding is expected to make aquatic habitats at MRA sites inimical for mosquitoes, however the tidal velocity and the degree of vegetation (thus affording shelter) will be important, particularly as some of the sites undergo higher rates of silt accretion and development of more complex habitats. This theory was tested using an additional site, an extant salt marsh site at Seymer's Marsh, on Brownsea Island (Dorset). A stepwise regression analysis showed a significant association with smaller water bodies and a vegetated substrate suggesting that over time, sites like Alkborough, where tidal velocity is low and where the potential for flooded vegetated habitat is high, might support a range of aquatic habitats for floodwater mosquitoes, even where there is regular tidal flushing. However, so far the numbers of aquatic habitats at this site for mosquitoes are minimal and could easily be contained.

This study focussed on eight MRA sites, varying in age from five to 16 years. The risk for potential future malaria transmission associated with these habitats is currently considered very low, as the degree of human biting by anophelines resulting from the MRA sites and the presence of humans infected with malaria within the vicinity of the MRA sites are both negligible. There is likely to be some nuisance biting associated with sites that support flooded habitats that are not regularly flushed by the tide, and a management plan may be needed to deal with these sites if they continue to promote nuisance biting by Ae. detritus and Ae. caspius. Similar controls would also be needed in the future if these species become implicated in disease transmission cycles, particularly as both Ae. detritus and Ae. caspius are identified as possible arbovirus vectors elsewhere in Europe (Medlock et al. 2005, 2007). Medlock et al. (2005) identified both species on account of their host-feeding preferences as possible WNV bridge vectors, and Ae. caspius has been implicated as a vector of Rift Valley fever virus in Egypt (Turell et al. 1996). Although the evidence for these species to become involved in arboviral transmission is limited, they may be a nuisance and therefore it is crucial that in the planning and design stages of future MRA sites, consideration is given to the creation of potential mosquito aquatic habitats, with further consideration to how such habitats could be minimized. Ongoing monitoring of such sites for mosquitoes should be instituted as part of regular faunal and floral monitoring and public health risk assessment for each site.