The Evacuation Of The Carteret Islands Biology Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

On the 22nd April 2009, the evacuation of the Carteret Islands began. Dubbed the 'first official climate change refugees', the 40 families who inhabited the tiny islands in the South Pacific were the first entire community to be displaced by the effects of climate change. As small as the population may have been, this was the unprecedented exodus of an entire community threatened by rising sea levels due to anthropogenic global warming. Rising sea levels threaten to completely submerge their homes and crops by 2015. Although it went almost unnoticed (Monbiot, 2009), this event foreshadows the tens to hundreds of thousands of people across the globe who may become permanently displaced by the middle of this century due to rising sea levels, intense droughts and heavier floods (Stern, 2007, Bindoff et al., 2007). With about 200 million people living in coastal floodplains (Milne et al., 2009), atmospheric warming of 3 - 4°C could result in the flooding of tens to hundreds of millions more people each year (Stern 2007). More than one fifth of Bangladesh could be under water with a sea level rise (SLR) of 1 m and large coastal cities such as London, Cairo, Tokyo and New York are at risk from rising sea levels (Stern, 2007). There is significant evidence to suggest that global mean sea level is rising significantly, after a period of little or no increase for almost 2,000 years (Bindoff et al., 2007).

Understanding changes in sea level, both on global and regional scales, is a difficult physical problem as a number of complex mechanism, which work over varying timescales, play a role (Rahmstorf, 2007) (see Figure 1). Climate-driven sea-level changes are caused by various interactions between components of the Earth system - predominantly oceans, ice sheets and the solid Earth (Milne, 2008). There is significant variability in the changes in sea level, both temporally and spatially. Global changes in sea level are caused by processes such as changes in the volume of water in the ocean basins (Church et al., 2001), changes in temperature of the ocean water and galcio-eustasy (changes in sea level due to storage or release of water from a glacier) (Milne, 2008). Regional changes in sea level can be caused by factors such as compaction of sediments, subsidence ice-water gravitational attraction, hydro-isostasy (loading and unloading of ocean basins) (Lambeck and Chappell, 2001) and by departure from the geoid (Barnett, 1983; Nicholls et al., 1999).

Figure 1. Schematic depiction of an ocean basin, showing processes that can cause relative and absolute sea-level change over various timescales (from seconds to millions of years) (from Milne, 2008).

Changes in sea level can occur from very short timescales, e.g. daily, to millennial timescales. Long-term changes in sea level can be due to subsidence of plate tectonics, changing the volume of ocean basins, or causing regional changes in sea level where the land is actually rising, e.g. in Scotland (Gehrels, 2010). Medium-term changes in sea level, e.g. over decades or centuries, tend to occur as a result of variations in ocean water density, caused by changing atmospheric temperature, ice melt, changes in ocean circulation (e.g. the Meridional Overturning Circulation) (Milne et al., 2009) and by anthropogenic or natural land-water storage. Steric contributions to sea level, caused by variations in ocean temperature (thermosteric) and salinity (halosteric) may be significant in some areas (Milne, 2008). Sea level will rise if the oceans warm and fall if it cools due to a change in the density of the water column; however, thermal expansivity, and hence global sea level, does not change linearly with global heat content, since water at higher pressure expands more than at lower pressure (Bindoff et al., 2007). Short timescale and extreme variation in sea level can be caused by weather events such as tsunamis and storm surges, the number of which have increased since 1975 (Bindoff et al., 2007).

Changes in sea level are measured in two main ways: 'relative sea level' or 'absolute sea level' (Milne et al., 2009, p.472). Relative changes in sea level are those relative to the ocean floor and are measured along the coast using tide gauges (Church et al., 2001) and can be caused either by movement in the land (either via isostatic or tectonic effects) or by changes in the height of the adjacent sea surface (eustatic changes in sea level) (Dorale et al., 2010). Absolute sea level is the level relative to the Earth's centre of mass (Milne et al., 2009) and can only be measured via satellite altimetry.

The most recent observations of changes in sea level are the most comprehensive ones (Milne et al., 2009). From 1992 onwards, satellite altimetry has provided near-global maps (66°S to 66°N) of absolute sea level at ten-day intervals, with an accuracy of approximately 5 mm (Nerem and Mitchum, 2001, cited by Bindoff et al., 2007). Altimetry measures the distance between the satellite and sea surface; measurements include variations in the geoid, caused for example by postglacial rebound (Peltier, 2006). These long-term variations can affect the estimate of mean sea level and therefore must be corrected for by accounting for glacial isostatic adjustment (GIA), estimated be approximately -0.3 mm yr-1 (Peltier, 2006). The importance of GIA correction to obtain accurate observations from satellite data can be seen from Figure 2. Since 2000, two additional observing systems have been implemented which compliment the altimetric data. The monitoring regime now includes a network of over 3000 Argo floats, which measure temperature and salinity in the top 1-2 km of the ocean (Milne et al., 2009) and the Gravity Recovery and Climate Experiment (GRACE) satellite mission, which takes monthly measurements of the global gravity field. These observing systems facilitate the separation of the contributions to changes in sea level from changes in ocean water density and changes in ocean mass (Milne et al., 2009). The global tide gauge network is useful for interpreting older measurements and is the primary source of information for deducing secular trends (i.e. longer than ten years) in sea level of the past century; however, it is a highly incomplete observing system (Milne et al., 2009) and is likely to contain both tectonic and eustatic signals (Church et al. 2001).

Figure 2. a) ocean mass changes from Grace (2003-2008). Upper curve is the raw data and the lower curve is the GIA-corrected time-series. b) Total ice sheet contribution to sea level estimate from Grace (2003-2008). The lower curve is the raw data and the upper curve is the GIA-corrected curve (from Cazenave et al., 2008).

Paeloclimatic data, based on geological observations, are very important for deducing past variations in sea level and linking them to the climate at the time. A better understanding of the Earth system responses and feedbacks which affect sea level facilitates more accurate modelling of present and future sea-level change (Gehrels, 2010). Evidence of past sea-level changes can be seen from morphological evidence and from stratigraphic or sedimentary records (Nichols, 1990; Hanebuth et al., 2000). It is useful to consider the response of ice sheets during past interglacial periods as ice extent was similar to that at present (Milne et al., 2009). During the last interglacial period, 125,000 years before present (125 ka), global sea levels were likely to have been between 4 and 6 m higher than in the 20th Century (Jansen et al., 2007). It is thought that during the Last Glacial Maximum (21 ka, LGM), global mean temperature was 4 - 7°C lower than at present, and sea level was about 120 m lower (Waelbroeck, 2002 and Schneider et al., 2006, cited by Rahmstorf, 2007). Various studies suggest that the rate of SLR during the Holocene (c. 12 ka - present) was between 0.1-0.2 m per century (van de Plassche et al., 1998; Gehrels et al., 2006; and Lewis et al., 2008 cited by Gehrels, 2010), with higher rates during the early Holocene than mid-Holocene (Fairbanks, 1989; Church et al., 2001). As stated by Gehrels (2010), throughout the Holocene, global patterns of changes in sea level have been controlled by regional sea-level variability, resulting from, for instance, a combination of isotatic and tectonic processes and steric effects. Geological observations suggest that sea-level change was at a rate of 0.0-0.2 mm yr-1 in the past 2,000 years (Jansen et al., 2007). Rapid changes in sea level were detected during the 20th Century (Gehrels et al., 2004), with two main phases in the rate of observed mean SLR: before 1940 and again after 1980, shown to be in close correlation with the global temperature evolution during that time (Rahmstorf, 2007). For the 20th Century, high-quality records from tide gauges demonstrate a rise of between 1.5 and 2mm yr-1 for the past 70 years (Peltier, 2001; Douglas, 2001, cited by Bindoff et al., 2007). A similar estimate of global SLR of 1.8 ± 0.3 mm yr-1 between 1950 and 2000 was made by Church et al. (2004).

Estimates of the current rate of SLR varying greatly depending on the period considered. There is still a lot of uncertainty regarding quantification of the sea level response to climate forcing and indeed, to the likely climate forcing of the coming decades; however, it is widely acknowledged that rising sea levels are a significant consequence of increased greenhouse gas emissions and global warming (Bindoff et al., 2007; Milne, 2008). Many studies have shown that observed changes in sea level and the rates of the change varied both spatially and temporally during the 20th Century (Milne et al., 2009). In some regions, rates are up to several times the global mean rise, while in some places sea level is falling, with a global average of SLR of 1.7 mm yr-1 during the 20th Century (Bindoff et al., 2007). Church et al. (2001) deduced an average SLR of ~1-2 mm yr-1 during the 20th Century from tidal gauges. Modelled estimate of the rate of sea-level change for this period was -0.8-2.22 mm yr-1, where the large uncertainty was due lack of knowledge regarding anthropogenic land water change (Bindoff et al., 2007).

In recent decades, rate of global sea-level rise has increased dramatically to a few decimetres per century; up from a rate of a few centimetres per century over the past few millennia (Milne et al., 2009). These variations have not been synchronous, which is likely to be a reflection of the spatial variability of sea level change due to the influence of land-ice changes, ocean temperature changes and long-period ocean dynamics (Milne et al., 2009). Considerable interannual and decadal variability in global ocean heat content (and hence sea level) has also been observed (Bindoff et al., 2007). Satellite altimetry has also identified areas where the rate of SLR is significantly more than the mean and areas (such as the north-eastern Pacific) where sea level has actually fallen since the commencement of satellite observation (Milne et al., 2009). As described in the Fourth Assessment Report from the Intergovernmental Panel on Climate Change (AR4, IPCC), rather than reflecting long-term trends, these spatial patterns probably reflect decadal fluctuation (Bindoff et al., 2007). A number of studies show that in certain locations, dynamical processes due to atmospheric forcing persist on multi-decadal timescales and may have contributed significantly to the observed 20th Century trend (Milne et al., 2009).

Since 1990 SLR has followed the uppermost uncertainty limit of the IPPC's Third Assessment Report (AR3) in 2001 (Rahmstorf, 2007). Using data from GRACE gravimetry, satellite altimetry and Argo floats, Cazenave et al. (2008) found that although sea level is still rising, since 1993 the rate of SLR has actually decreased from the previous decade. Since the IPCC published the estimate of 3.1 mm yr-1 for the decade 1993-2003 (shown in Figures 3 and 4), ocean thermal expansion change, shown from Argo float data, showed a plateau even though SLR was still occurring, but at a reduced rate of ~2.5 mm yr-1 (Cazenave et al,. 2008). Large spatial variation in this thermosteric signal was observed during this period (Milne, 2008).

Figure 3. Variations in global mean sea level between January 2003 and October 2005 (compared with the mean 1993 to mid-2001). The blue solid line corresponds to 60-day averaging of the data (from Bindoff et al., 2007).

Figure 4. Variation in global mean sea level due to thermal expansion, 1955-2003. The black and red lines denote the deviation from the 1961-1990 average. The green line shows the deviation from the average of the black curve for the period of 1993-2003 (from Bindoff et al. 2007). The shaded area and the error bars represent the 90 percent confidence level.

There is significant evidence to suggest that current regional and eustatic SLR is a consequence of climate change, resulting in melting of land ice and the thermal expansion of ocean water (Bindoff et al., 2007; Stern, 2007; Milne et al. 2009; Vermeer and Rahmstorf, 2009). Many studies have concluded that thermosteric effects have contribute significantly to recent changes in sea level and it is thought that it will be the dominant contributor to SLR over the 21st Century as global atmospheric temperatures continue to rise (Meehl et al., 2007). The input of large volumes of freshwater into the oceans from melting ice sheets and glaciers has a significant potential effect on changes in sea level (Rahmstorf, 2007).

The above-mentioned IPCC estimate of the rate of global mean SLR of 3.1 mm yr-1 between 1993 and 2007 is thought to be attributable in equal amounts to ocean-water density changes (owing to thermosteric and halosteric changes) and due to the contribution of freshwater from melting of land ice (Bindoff et al., 2007). A more recent (altimetry-based) analysis of SLR between 2003 and 2008 from Cazenave et al. (2008) was for a contribution of 20 percent from thermal expansion and 80 percent from land-ice melt. The findings of the study by Cazenave et al. (2008) indicate that approximately half of the ocean mass increase since 2003 is from the enhanced contribution of the polar ice sheets (enhanced in comparison to the findings of the previous decade) and half from the melting of mountain glaciers. These contributions equate a 2 mm yr-1 ocean mass contribution between 2003 and 2008, i.e. 80 percent of the altimetry-based rate of sea-level change during that period (Cazenave et al., 2008). Cazenave et al. (2008) found the estimates of the steric sea level contribution from both Argo float data (~0.3 mm yr-1 between 2003 and 2008) and from the difference between altimetry-based sea level and ocean mass change (0.37 mm yr-1 between 2004-2008) to be in good agreement.

Significant observational evidence suggests that there has been a reduction in volume of ice caps and glaciers over the past 20 years as a result of changes in temperature and precipitation (Meehl et al., 2007). The rate of melting is rapid and enhanced by the positive feedback associated with the surface energy balance associated with exposing land surfaces and shrinking glaciers (Meehl et al., 2007). The estimates in the contribution of the ice sheets vary significantly. The Scientific Committee on Antarctic Research (2009) estimate that around 2005, the Antarctic Peninsula contributed to global SLR at a rate of 0.16 ± 0.06 mm yr-1, which is a less than the IPCC's estimated contribution of 0.21 ± 0.35 mm yr-1 to SLR between 1993 and 2003. Cazenave et al. (2008) estimate that the net Antarctica contribution at 0.54 ± 0.2 mm yr-1 for 2006, three times Meier et al.'s estimate of 0.17 mm yr-1. With increasing global atmospheric temperatures scientists believe that acceleration in glacier loss over coming decades is likely (Meehl et al., 2007; Pfeffer et al., 2008), possibly up to a volume loss of 60 percent by 2050 (Schneeberger et al., 2003, cited by Meehl et al., 2007).

Rapid melting of large polar ice sheets lead to distinct and highly non-linear variation in the global distribution of meltwater, which have been termed 'sea-level fingerprints' (Clark et al., 2002; Tamisiea et al., 2003). The concept of sea-level fingerprinting provides geographic explanations in variations in sea level by determining where past ice melt sources were located (Tamisea et al., 2003; Gehrels, 2010). The physics of the fingerprints have been explored in detail by Tamisiea et al. (2003) by splitting the sea-level change into contributions from radial perturbations in the two bounding surfaces, i.e. the sea surface (or geoid) and the solid surface. Tamisiea et al. (2003) described that sea-level fingerprinting, in the event of significant melting, is characterized by a drop in the sea level in the near-field of the ice complex and a gradually increasing rise in sea level at distances from the ice sheet; quantified as being 0.0 - 1.3 times the eustatic SLR, progressively increasing with distance from the ice sheet (up to ~90°) (see Figure 5). The fall in sea-level near to the ice sheet is a result of the relaxation of the sea-surface as the gravitational pull of the ice sheet weakens, as well as a result of ocean-plus-ice unloading of the solid surface (Tamisiea et al. 2003). The far-field increase in sea level is largely driven by this relaxation.

Figure 5. Schematic representation of the near- and far-field sea-level change resulting from diminishing gravitational attraction of shrinking ice mass (from Tamisea et al., 2003).

An example of fingerprinting approach applying such corrections is the study by Clark et al. (2002), who deduced that the Laurentide Ice Sheet could not have been the sole contributor to the meltwater pulse 1A (between 13 and 14.6 ka), but there also appears to have been substantial contribution from the Antarctic Ice Sheet. There are other examples where fingerprinting has been used to estimate that the melting of the Greenland Ice Sheet contributed 1.0±0.6 mm yr-1 to global SLR since 1960 (Marcos and Tsimplis, 2007). Fingerprinting is thus considered a useful methodology for deducing the source of contributions to sea-level changes.

The IPCC AR4 included a table which summarised the total budget of global mean sea level change over 1961-2003 and 1993-2003 and estimated the contributions to the budget, presented in Table 1 and Figure 6. Omitted from this table are contributions known to be small, such as changes in atmospheric water vapour and climate-driven changes in land water storage, as well as the contribution from changes in anthropogenic land-water storage since large uncertainty remains in quantifying this contribution (Bindoff et al., 2007).

Table 1. Estimates of the upper four entities contributing to the budget of global mean SLR for 1961-2003 and 1993-2003, compared with the observed rate of rise. Ice sheet mass loss of 100 Gt yr-1 is equivalent to 0.28 mm yr-1 SLR (from Bindoff et al., 2007).


Sea Level Rise (mm yr-1)


1961 - 2003

1993 - 2003

Thermal Expansion

0.42 ± 0.12

1.6 ± 0.5

Glaciers and Ice Caps

0.50 ± 0.18

0.77 ± 0.22

Greenland Ice Sheet

0.05 ± 0.12

0.21 ± 0.07

Antarctic Ice Sheet

0.14 ± 0.41

0.21 ± 0.35


0.1 ± 0.5

2.8 ± 0.7


1.8 ± 0.5

3.1 ± 0.7


Difference (sea level enigma)

0.7 ± 0.7

0.3 ± 1.0

Figure 6. Estimates of the various contributions to the budget of global mean sea level change (upper four entities) for 1961-2003 (blue) and 1993-2003 (brown). The bars represent the 90 percent error range.

Interestingly, as can be seen from Table 1 and Figure 6, there is a discrepancy between the observed SLR and the addition of all of the known contributions to the rise, where the observed rise is slightly more (Gehrels, 2010). This has been referred to as 'sea-level enigma' by Munk (2002, cited by Gehrels, 2010) or the 'attribution problem' by Miller and Douglas (2004, cited by Gehrels, 2010). This lack of closure in the sea level budget is significant - for example, in the IPCC's AR4, this shortfall in the sum of the contributions compared with the observed changes in SLR was 0.3 ± 0.1 mm yr-1 for the period 1993-2003, i.e. approximately 10 percent. This was a decrease in the discrepancy between the sum and observed values in the IPCC's AR3 in 2001 (Gehrels, 2010). According to Gehrels (2010), this inability to close the sea level budget implies significant uncertainties in the calculations, stemming either from an instrumental overestimation of true global seal-level change, an underestimation of the contributions or from the existence of unknown sources of sea-level change.

For time scales relevant to anthropogenic warming, Vermeer and Rahmstorf (2009) propose that the rate of change of sea level is proportional to the magnitude of the global atmospheric temperatures above those of the pre-Industrial Age. This is not in keeping with the IPCC AR4 (Bindoff et al., 2007), where the estimate for the rate of SLR by 2100 using various global temperature scenarios was in the range 0.75 - 1.90 m between 1990 and 2100. The limitation of the latter estimate, as pointed out by Vermeer and Rahmstorf (2009), is that it does not include rapid ice flow changes; it was argued at the time that they could not be modelled accurately. The estimate thus did not present an upper limit of the expected rise, provoking many attempts of semi-empirical approaches to projecting future sea-level changes (Bindoff et al. 2007), most recently by Vermeer and Rahmstorf in late 2009 (see Figure 7). This approach includes a rapid response term, which accounts for some components of sea level adjusting quickly to changes in temperature, e.g. the heat content of the surface mixed layer (Church et al., 2001; Vermeer and Rahmstorf, 2009). In this study, Vermeer and Rahmstorf concluded that by 2100, sea levels could actually rise by approximately three times that estimated within the IPCC's AR4 (excluding rapid ice flow dynamics) (Figure 8). Sea level could rise even faster yet, due to massive uncertainties in making the predictions.

Figure 7. Semi-empirical approach to estimating sea-level rise during the instrumental period. (Upper) Observations-based rate of SLR - red: with tectonic and reservoir effects removed, gray: predicted using a semi-empirical approach which includes the rapid response term and blue (with uncertainty estimate): predicted using a semi-empirical approach (without a rapid response term) using observed global mean temperature data. (Lower) The integral of the curves in the upper plot, the thin red line shows the annual sea-level values, which is shown to be almost obscured by the dark blue line, which shows the predictions from the semi-empirical approach which includes the rapid response term (from Vermeer and Rahmstorf, 2009).

It is thought that land ice will continue to lose mass increasingly rapidly during the 21st Century (Bindoff et al. 2007) although it is not known whether discharge from ice sheets will continue to increase as a consequence of accelerated ice flow. Melting of smaller ice caps and mountain glaciers may cause up to 0.25 m of SLR by 2100, in addition to SLR from the two major ice sheets (Meier et al., 2007), the melting of which could raise sea levels but up to 7 m (Meehl et al., 2007). A study by Pfeffer et al. (2008) concluded that due to geological constraints, sea level is very unlikely to increase by more than 2 m by 2100. This study was conducted on the basis of climate modelling and analogies with past conditions and showed that more plausible conditions would increase sea level by approximately 0.8 m, rather than the 2 m caused if all variables are quickly accelerated to extremely high conditions.

Figure 8. Projection of SLR from 1990 to 2100 (temperatures used based on IPCC temperature projections for three different emissions scenarios). The SLR projected by the IPCC in the AR4 for these emissions are shown for comparison in the bars on the bottom right. The red line shows the observations-based annual global sea level data (from Vermeer and Rahmstorf, 2009).

Spatial non-uniformity in the changes in sea level caused by climate change is a serious issue for considering the hazard of future SLR (Milne, 2008). Over the coming century, some regions are likely to experience considerable increases in sea levels, whereas sea level change may be negligible in some areas and a considerable fall in levels may even be observed in others (Bindoff et al., 2007, Milne, 2008). Increasing sea levels have considerable environmental, societal and economic implications; including increased coastal flooding, loss of wetlands and wildlife habitat, displacement of millions of people, increased cost of coastal protection, decreased freshwater availability due to saltwater intrusion (Bindoff et al., 2007; Stern, 2007).

There are significant uncertainties regarding future sea level predictions (Bindoff et al., 2007; Hunter, 2009; Milne et al., 2009; Vermeer and Rahmstorf, 2009; Gehrels, 2010), attributable to uncertainties in ice flow dynamics (Milne et al., 2009; Gehrels, 2010), contribution of terrestrial water sources and anthropogenic water storage (Nerem, 2006), contributions of land ice due to lack of information regarding the total amount of ice and the rate at which it is likely to melt (Nerem, 2006). Uncertainties in quantifying past sea level changes has direct consequences for model predictions - the spatial variability of the likely changes in sea level is a very difficult problem (Milne et al., 2009). Another huge source of uncertainty stems from the predictions of likely changes in global atmospheric temperatures for 21st Century, which directly influence the change in sea-level (Vermeer and Rahmstorf, 2009). The future social, economical and technological development of the world is also not known (Hunter, 2009).

For example, there is great uncertainty regarding the magnitude of the contributions of the Greenland ice sheet, West Antarctic ice sheet and mountain glaciers in the Himalayas and Andes to SLR in response to global warming (Carlson et al., 2008; Scientific Committee on Antarctic Research (SCAR), 2009). It has been suggested that there is a 5 percent probability that the melting of the West Antarctic ice sheet could cause a SLR of 10 mm yr-1, and a 30 percent chance that it will cause SLR at a rate of 2 mm yr-1 (SCAR, 2009). Figure 9 shows the modelled SLR by 2100 with meltwater contributions from ice sheet and mountain glaciers.

Figure 9. (a) Sea-level change predicted by assumed 1 mm yr-1 melt from Antarctica. (b) Sea-level change predicted by assumed 1 mm yr-1 from Greenland. (c) Sea-level change predicted by assumed 1 mm yr-1 melt from mountain glaciers. (from Gehrels, 2010).

Progress in decreasing uncertainties and improving monitoring and modelling is rapid. As longer time-series with the monitoring combination of satellite altimetry and global tide gauge data, interannual variability will become less dominant and it will be easier to isolate the causes of both regional and decadal variability (Milne et al., 2009).

It is widely acknowledged that rising sea levels are a hazard associated with increasing global temperatures (Bindoff et al., 2007; Stern, 2007; Milne, 2008; Vermeer and Rahmstorf, 2009). It can be concluded that, although regional sea levels fluctuate considerably, significant evidence exists that shows that global mean eustatic sea level is rising, and is predicted to continue rising during the 21st Century. Global mean sea level was observed to have increased by approximately 3 mm yr-1 over 1993-2007 (Bindoff et al., 2007). The main causes of SLR are land ice melt and ocean water density changes. It is thought that at present, the dominant factor influencing SLR is the melting of land ice. Significant uncertainties remain in estimating the cause of sea level changes and in making predictions for likely future changes. Estimates of up to 2 m rise in sea levels from 1990 values have been made (Pfeffer et al., 2008). Improved and better-constrained estimates of the projected rate of SLR are continually being made. It is important that future research focuses on gaining better understanding of the oceanographic and climatic processes which are influencing temporal and spatial variability in sea-level changes. Accurately projecting future SLR is thus a priority so that planning of mitigation methods or adaption methods, such as coastal defences, can be implemented to minimise the hazard.