Human migration and agricultural development due to sea-level rise likely to cause biodiversity loss

Previous studies of the impacts of sea level rise (SLR) on human inhabitance patterns and terrestrial biodiversity have focused mainly on primary effects, which include land area loss due to inundation and erosion in coastal areas. However, secondary effects, or the ecological impacts of human displacement and relocation from low-lying areas because of primary effects, are very important when considering biodiversity loss. Wetzel et al. (2012) examined the impacts of secondary effects on biodiversity in the Southeast Asian and Pacific (SEAP) regions and found that in many cases secondary effects may dominate range loss. In the predicted levels of SLR, 4–27% of the human coastal population (2-52 million people) will be forced to relocate. Human migration will be especially pronounced on the Indo-Malaysian islands, where 4–28% of the population will be forced to migrate, and 30% of the inundated land will be urban and intensive agricultural land. This relocation, as well as the increase in natural land conversion to agricultural and urban land in the hinterlands will likely cause significant population declines in many species.–Olivia Jacobs
Wetzel, F.T., Kissling, D., Biessmann, H., Penn, D.J. Future climate change driven sea-level rise: secondary consequences from human displacement for island biodiversity. Global Change Biology 18, 2707–2719. [GSSS Wetzel kissling secondary]

                  Wetzel et al.(2012) studied the potential effects of sea level rise (SLR) on more than 1,200 islands in the Southeast Asian and Pacific region (SEAP). They further divided this area into three broad categories, Australia, Oceania, and Indo-Malaysia, and used Digital Elevation Model (DEM) data to map potential SLR scenarios. The three scenarios addressed were 1 m, 3 m, and 6 m of total SLR, which are values that represent the most common estimates of SLR for this century and higher ones for the coming centuries. Using these maps of SLR, the scientists analyzed primary effects on coastal areas by estimating the inundated and eroded area in the coastal zones. Secondary effects, or the land converted from natural habitat to agricultural or urban areas due to population migration, were estimated by using mammal distribution data for 54 species on 109 Indo-Malaysian islands. Each mammal was analyzed individually using information from previous studies. Notably, the scientists also assumed that there was an equal-area land-conversion of inundated urban and intensive agricultural areas in the hinterlands.
                  The data indicate that primary effects will result in a large loss of coastal zone in the SEAP region. On average, 3% of land area in the SEAP will be inundated from the 1 m scenario, and 32% will be lost in the 6 m scenario. This will cause an estimated 8–52 million people (or 4–27% of the total population) to migrate to the hinterlands. Further, the Oceanic area will be most vulnerable to inundation, with an estimated 7–46% loss for the 1–6 m scenarios. Australia will be less affected, experiencing a land area loss of 2–25% in these scenarios, and Indo-Malaysia will experience a land loss of 4–35%. In these inundated areas, about 30% of the lost area is urban and intensive agricultural land in Indo-Malaysia and 4–28% of the population will be forced to migrate, whereas only 2–6% of the population in Oceania and Australia will be at risk. These data indicate that biodiversity loss due to secondary effects will be most pronounced in Indo-Malaysia and primary effects will dominate elsewhere.
                  Secondary effects are especially important when considering biodiversity because areas with large coastal urban populations and agricultural development, such as Indo-Malaysia, will experience large amounts of human migration and land conversion. Range loss due to secondary effects was greater than loss due to primary effects in Indo-Malaysia for 22–46% of species in maximum range loss scenarios, while 9% of species are only vulnerable to secondary effects. The authors also note that different species in similar island scenarios respond very differently to primary and secondary effects. The Smokey Flying Squirrel has a projected range loss of 1–3% due to primary effects, but a loss of 2–60% from secondary effects. Comparatively, the Rajah Sundiac Maxomys may experience primary loss of 1–15%, while secondary effects play a lesser role in range area loss. Thus, in some instances secondary effects will dominate species range loss, while secondary effects will not contribute significantly to that of other species. The effects are dependent on both region and species.
                  This study makes some key assumptions that could contribute significantly to a change in the aforementioned estimates. It ignores the expected temperature and precipitation changes due to climate change, and it also only analyzes the secondary effects on mammals on the SEAP islands. The model may also be overly conservative because it only considered the most intensive farming types, did not consider other ecological interactions such as interspecific competition due to coastal species migration, and did not include the worst-case and more liberal estimates of SLR. However, the study still provides further evidence that SLR anticipated from global warming will have many major consequences for biodiversity, and it expands our understanding of biodiversity loss to include the impacts of secondary effects due to SLR.

Sea Levels and Ocean Thermal Expansion

Data show that sea levels have been rising significantly over the past half century. The contributions from thermal expansion of the oceans, the melting of glaciers, and loss of ice masses in Greenland and Antarctica are commonly studied, but together they do not account for the total sea level rise (SLR). The rate of SLR is ~1.8 mm yr–1, but the total contributions from these sources is estimated to be about ~1.1 mm yr–1, which leaves ~0.7 mm yr–1 unaccounted for. Pokhrel et al. (2012) examined terrestrial water storage, namely reservoir water impoundment and unsustainable groundwater irrigation, and found that these types of sources likely contribute about 0.77 mm yr–1 to SLR. Thus, anthropogenic water uses contribute greatly to changing sea levels, and can help close the gap in the global sea level rise.¾Olivia Jacobs
Pokhrel, Y.N., Hanasaki N., Yeh, P.J., Yamada, T.J., Kanae, S., Oki, T. Model estimates of sea-
level change due to anthropogenic impacts on terrestrial water storage. Nature
Geoscience 5, 389392.

Pokhrel et al. (2012) used an integrated water resource assessment modeling framework to analyze the potential contributions of anthropogenic sources to global sea level change (SLC). They estimated many values such as the amount of water lost to seepage in reservoirs, the number of years to fill a reservoir, and the rate of change in loss over time, because no real values are currently available. Using these estimates and data from numerous supplementary sources, they concluded that the global reservoir capacity is about 8,000 km3. They also estimated unsustainable groundwater use based on total water demand and the availability of water from near-surface sources, which they concluded gave them an accurate estimate of groundwater contributions to SLR.
                  By comparing their estimated values with other model simulations, Pokhrel et al.were able to estimate the relative contributions from different types of terrestrial water storage (TWS) to SLC. They focused on artificial reservoir impoundment and unsustainable groundwater use and found that these two variables can close the gap in SLC quite successfully. Artificial reservoirs generally cause a drop in the sea level by holding water over land, and this study estimated that the cumulative contributions of reservoir capacity and storage to SLC was ~22 and ~15 mm, respectively. Further, when seepage from reservoirs was accounted for, the estimates of artificial reservoir capacity was ~31 mm, and the actual estimated contribution was ~21 mm. In both of these scenarios, the scientists noted a large discrepancy between reservoir capacity and reservoir storage.
Unsustainable groundwater use contributes to SLR because water is removed from the ground and eventually ends up in the oceans. Here, the scientists estimated groundwater depletion (GWD) has contributed ~48 mm to cumulative SLR. Also, climate-driven terrestrial water storage (TWS), including soil moisture and snow and river storage exclusive from Greenland and Antarctica, has a net contribution of ~8 mm to global SLR. Thus, while climate-driven TWS has significant decadal and annual variation, the long-term contributions to global SLR are relatively small and anthropogenic redirecting of water contributes to SLR much more significantly. Pokhrel et al. also considered the uncertainty in their estimates, and found that the uncertainty of net TWS, including groundwater depletion, climate-driven TWS, and reservoir storage, could be as high as 30%.
Overall, the annual estimated contribution of these TWS sources to SLC was +0.77 mm yr–1, which is close to the previously unexplained gap in SLR of 0.70 mm yr–1. Individually, the estimates were +1.05 mm yr–1 from groundwater, +0.08 mm yr–1 from climate-driven TWS,       –0.39 mm yr–1 from reservoir storage, and +0.03 mm yr–1 from the Aral Sea, which was a main source of water diversion for irrigation. Comparatively, the contribution from thermal expansions of the oceans was ~0.42 mm yr–1, ~0.5 mm yr–1came from glaciers and ice caps melting, and ~0.19 mm yr–1 from ice-mass loss in Greenland and Antarctica.
The trends in TWS over the past 50 years indicate that groundwater depletion has been increasing significantly over time and may increase more in the future, and the climate-driven TWS has also been increasing in recent years. Reservoir impoundment has also recently leveled off, so there may be an even greater contribution to SLR from TWS sources in the future, and while there is obvious uncertainty in estimations used in this study, they are all within the plausible limits for many countries. The scientists note that some sources, such as the effects of deforestation and wetland drainage, were not considered, but these sources have previously been shown to contribute very little to global SLR. Thus, these TWS sources can help close the gap in estimations of global sea-level change, and can help shed light on the future forecasts for global SLR.

Corals in Tahiti Shed Light on the The Bølling Warming Period and Sea Level Rise

The Bølling Warming period was the last deglaciation period and occurred about 14,650 years ago. During this time, global sea levels rose about 20 meters in less than 500 years. The source of this meltwater pulse (MWP-1A) and its temporal and causal relationship to the abrupt climate changes of this period remain unclear, but Deschamps et al. (2012) were able to further our understanding of this event. Using coral samples taken from the reefs in Tahiti, the scientists identified a probable total sea level rise between 14 and 18 meters during this time period and an annual sea level rise of 40 millimeters per year. Further, contrary to previous studies, this study concluded that the Antarctic Ice Sheet contributed significantly to global sea level rise during this period. Ultimately, understanding the dynamics of this warming period help scientists understand current warming patterns, as ice sheets today contribute significantly to the global sea level rise and could therefore influence future warming patterns.¾Olivia Jacobs
Deschamps, P., Durand, N., Edouard B., Hamelin, B., Camoin, G., Thomas, A.,
Henderson, G., Okuno, J., Yokoyama, Y. Ice-sheet collapse and sea-level rise at
the Bølling warming 14,600 years ago. Nature 483. 559–564. [GSSS Deschamps
Durand Bolling]

                  Deschamps et al.(2012) collected coral samples from three distinct areas in reefs near Tahiti during the Integrated Ocean Drilling Program (IODP) Expedition 310. Using U-Th dating and knowledge of the ranges of modern reef environments the scientists were able to map previous sea levels and calculate the change in sea level over the past 16 kyr.
                  The oldest samples in this study were shallow water coralgal assemblages (± 0.03 kry BP and 16.09 ± 0.04 kry BP. These samples were therefore from before the onset of the Bølling Warming period, and they indicate a Relative Sea Level (RSL) of 117–107 meters below present sea level (m.b.s.l.). The IODP data also revealed a major discontinuity in the Tahiti RSL between 14.28 ±0.02 kry BP and 14.31 ± 0.04 kry BP, thus indicating the start of MWP-1A.
From these and similar data Dechamps et al. inferred a conservative estimate for post-MWP-1A of 88–83 m.b.s.l., which indicates a sea level rise of 14–18 meters during this warming period. The longest possible duration of this sea level jump is about 350 years, which equates to an average sea level rise of about 46 ± 6 mm yr­–1 in Tahiti. However, because of uncertainties in the data, the duration of this rise could have been much shorter, and so this estimate is considered a minimum value. 
Scientists originally believed that the sea level rise during MWP-1A was caused by partial melting of the Northern Hemisphere ice sheets (NHIS). This would have occurred if a large amount of fresh water were deposited into the North Atlantic, thereby slowing the Atlantic meridional overturning circulation (AMOC) and abruptly ending the Bølling warming period about 14.1 kyr ago. This scenario agrees with previous data in Barbados, which indicate a much later onset of the warming period. However, this study, and other data from reefs in Hawaii and elsewhere, indicate that the Bølling period happened about 500 years earlier than previously believed and it is therefore more likely that the sea level rise coincides with the inception of the Bølling period. Further, the Antarctic Ice Sheet (AIS) probably contributed significantly to the MWP-1A. While uncertainties in the data make it difficult to estimate the relative contribution of the AIS and NHIS to the total SLR, it is probable that the AIS contributed at least half of the eustatic sea level rise during this time period.
                  These new findings necessitate further investigation, as the initial trigger of the Bølling period is still unknown. However, one of two scenarios is plausible. In the first scenario, the rapid melting of the AIS triggered the intensification of the AMOC, which would have started the Bølling warming in the Northern Hemisphere and melted northern ice sheets. In the second scenario, the AMOC increased first and caused the warming, which led to the melting of the NHIS and subsequently collapsed the AIS. These two scenarios are not mutually exclusive either, as they could have acted together and reinforced one another.
While much is still unknown, the data in this study indicate that the Bølling period occurred much earlier than previously believed, and the AIS contributed greatly to this period of rapid sea level rise. Understanding more about this event may shed light on the dynamics of the current warming period, as ice sheets today contribute greatly to current sea level rise. 

The Contribution of Glaciers and Ice Caps to Sea Level Rise

The contribution of glaciers and ice caps (GICs) to global sea level rise is significant. However, most studies focus on the Greenland and Antarctic glaciers and their impact on sea level rise (SLR) while GICs for areas such as the Himalayas, Alaska, and the Alps are usually interpolated from more sparse mass balance measurements. This study used a combination of Gravity Recovery and Climate Experiment (GRACE) satellite data between 2003 and 2010 and analyzed global ice-mass changes in GICs. The data revealed that mass loss in GICs excluding Greenland and Antarctica was about 30% smaller than previous models predicted, or around 148 ± 30 Gt yr­–1. This equates to a sea level rise of about 0.41 ­­± 0.08 mm yr­–1 (Jacob et al. 2012). Although the data agreed with other measurements of sea level rise (SLR), the scientists note that the short study period, along with the high interannual variability during this time, should therefore encourage further investigation.¾Olivia Jacobs
Jacob, T., Wahr, J., Pfeffer, W.T., Swenson, S. 2012. Recent contributions of glaciers and ice caps to sea level rise. Nature 482, 514-518. 

                  Jacob et al. (2012) used Gravity Recovery and Climate Experiment (GRACE) satellite images to calculate the mass changes over all ice-covered regions greater than 100 km2across the world. They used data between 2003 and 2010, and divided 20 glacial and ice-covered regions into 175 smaller ‘mascons’ and evaluated the mass changes in each small area to analyze local changes. The areas studied included Iceland, High Mountain Asia, Alaska, Northwest America, Baffin Island, Scandinavia, Patagonia, and more. While Greenland and Antarctica were not the focus of this study, Jacob et al.used data from these areas to estimate the total contribution to sea level rise (SLR) from glaciers and ice caps.
                  By taking more accurate measurements of GICs across the world, Jacob et al. found that the mass balance changes in these regions was about 30% smaller than other studies had calculated. Globally, the GIC mass balance excluding Greenland and Antarctica was 148 ± 30 Gt yr­–1, contributing 0.41 ­­± 0.08 mm yr­–1 to SLR between 2003 and 2010. There have not been other estimates over this exact time frame, but similar studies indicate changes of 1.41 ±0.3 mm yr­–1 between 2001 and 2005 and 0.98 ± 0.19 mm yr­–1 between 2001 and 2004.
                  Jacob et al.focused attention on the High Mountain Asia (HMA) GICs as these calculations were significantly different than a recent GRACE study between 2002 and 2009. Previous data predicted a change of –55 Gt yr­–1 over the entire region and –29 Gt yr­–1 over the eastern Himalayas alone, but the data from this study predicted a change of –4 ± 20 Gt yr­–1 across the HMA region. To verify it, Jacobs et al. ruled out many other factors that could have given a low rate of loss. They considered groundwater loss by analyzing smaller mascons in the HMA plain region and found that loss happened at a rate consonant with their findings. They also considered tectonic processes as contributors to the low predicted rates of loss, and calculated that crustal uplift would have to be occurring at a rate of about 1 cm yr­–1 to account for their data, whereas Global Positioning Systems indicate long-term uplift rates of 0.5–0.7 cm yr­–1.
Jacob et al. thought it might also be possible that meltwater in the HMA might be sinking into the ground and remaining in the area, which would also create no change in GRACE imaging. However, local storage capacity in the HMA region is small due to thick permafrost layer, and thus it is unlikely that water could be stored here. Water may also be diverted for irrigation, but the irrigated areas lie well outside the HMA mascons, and thus GRACE would show a mass loss even if this water were being directed toward irrigated areas. By ruling out these potential factors of causation, Jacob et al. were able to further verify their findings and support the lower rate of loss in the HMA region.
Although this study focused on GICs rather than Greenland and Antarctica, Jacob et al. also compiled data from Antarctica and Greenland to calculate a total SLR contribution of 1.48 ± 0.26 mm yr­–1 between 2003 and 2010. This value compares well with other estimates of total SLR, but regional GIC data varied greatly between estimates and the scientists suggest that these differences should be further studied.
                  Additionally, this study was done on a very short time scale, and thus the results should be heeded cautiously. Interannual variability (due to seasonal changes) was very high during this time period, especially in regions such as High Mountain Asia, and Jacob et al. therefore suggest that this type of study should be extended to longer time periods in order to get a more accurate prediction of GICs and their impact on global SLR. 

Warmer Ocean Currents Melt Antarctic Ice Shelves and Causes Thinning

Ice shelves are thick, floating platforms of ice that extend from a grounded glacier into the sea. They help buttress the flow of grounded tributary glaciers into the ocean and therefore help slow global sea level rise. Eighty percent of grounded ice in Antarctica drains through ice shelves, and thus the thinning of ice shelves could lead to runaway ice sheets and an acceleration of melt patterns. On the Antarctic Peninsula, ice-shelf collapse has already led to a retreat and acceleration of several local glaciers. By mapping the glacier and ice-shelf elevation change, sea temperatures, and ocean trough depths, Pritchard et al. (2012) concluded that in 20 of 54 Antarctic ice shelves, ice-shelf thinning and glacial acceleration were driven by rapid basal melt. In all cases, tributary glaciers were also thinning, and accounted for about 40% of Antarctic discharge and the majority of ice-sheet mass loss. The scientists found that these areas were concentrated in West Antarctica, where changes in sea currents have brought warmer waters to the coast and into deep bathymetric troughs and melted the ice from below.¾Olivia Jacobs
Pritchard, H.D., Ligtenberg, S.R.M., Fricker, H.A., Vaughan, D.G., van den Broeke,
                  M.R., Padman, L. 2012. Antartic ice-sheet loss driven by basal melting of ice
shelves. Nature 484, 502–505. 

                  Pritchard et al. used satellite and laser altimetry to understand patterns of ice-shelf thinning and basal melt. They used laser altimetry to measure the surface height change in major Antarctic ice shelves between 2003 and 2008, and compared these changes to measured values of sea temperature. Data showed that thinning was strongly regional and was most rapid along the Amundsen and Bellingshausen Sea coasts.
During this time period, thick ice shelves thinned while thinner ones showed no significant pattern of further thinning in the areas near the Amundsen and Bellingshausen Sea coasts in West Antarctica. Firn modeling indicated that firn layers thickened throughout this region because of increased accumulation. Also, patterns of ice-shelf retreat and glacier influx were similar throughout different ice shelves, and so Pritchard et al. concluded that the changes in ice-shelf elevation were driven by something other than local climate. When these changes were mapped with regional sea temperature data, the data showed a strong correlation between thinning ice shelves and warmer water temperatures, indicating that basal melt was driving elevation changes.
The highest melt rate, 40 m yr –1, was near the grounding line of Pine Island Glacier in West Antarctica. In this glacier, the flow rate of grounded ice increased by 43%, and contributed to a global sea level rise of 1.2 mm per decade. As a whole, Pritchard et al. found that the most rapidly thinning ice shelves occurred in areas with high sea-floor temperatures and deep bathymetric troughs that spanned continental shelves in West Antarctica and ice shelves that did not show significant levels of thinning in West Antarctica were removed from these areas.
The warmer temperatures observed by Pritchard et al. were driven by fluctuating incursions of Circumpolar Deep Water (CDW). This change in ocean current is wind driven, saline, relatively warm, and dense, and, as this study showed, causes basal melting in ice shelves. The fluctuating CDW is channeled at depths below 300 m, and thus areas where bathymetric troughs approach the continental shelf are particularly susceptible to fluctuations in CDW. Many major glaciers that reside above these troughs are already in retreated positions and displayed greater rates of thinning.
Wind forcing patterns are not yet well understood, but it is clear that they have helped drive CDW shifts and are influenced by tropical Pacific sea surface temperature changes, ozone loss, and increased greenhouse gases. These shifting wind patterns also influence atmospheric temperatures, which drive surface ice loss in other ice shelves that are removed from the CDW on the Antarctic Peninsula. Thus, atmospheric changes influence melting above and below Antarctica’s ice shelves. Pritchard et al. conclude that these changing oceanic conditions have already driven profound changes in the ice sheets, and may have already triggered unstable glacial retreat and subsequently a substantial increase in global sea levels. 

Eemian interglacial reconstructed from a Greenland folded ice core.

The Eemian interglacial period occurred between 130,000 and 115,000 years ago. Previous attempts to extract an ice core from this time period have been unsuccessful, but scientists in Greenland have recently been successful. The North Greenland Eemian Ice Drilling (NEEM) team, extracted an ice core from the Northern Greenland ice sheet and were able to analyze atmospheric changes during this time using air bubbles frozen in the ice. Using stable water isotopes, the data reveal that surface temperatures at NEEM peaked at 8 ± 4°C above the mean of the past millennium during the Eemian interglacial period, followed by gradual cooling (Dahl-Jenson et al. 2013). The scientists also found that between 128,000 and 122,000 years ago the thickness of the ice decreased by 400 ± 250 meters, reaching surface elevations of 130 ±300 meters below present levels. Furthermore, melting and movement patterns of ice before the Eemian interglacial period indicate that significant melting and refreezing events occurred during these years.¾Olivia Jacobs
Neem Community Members, 2013. Eemian interglacial reconstructed from a Greenland
folded ice core. Nature 493, 489–494. 

            At the North Greenland Eemian Ice Drilling (NEEM) site, Dahl-Jenson et al.extracted a 2,540-meter ice core from the Greenland glacier between 2008 and 2012. Using this ice core, they were able to extract air bubbles trapped inside and analyze them based on their isotopic properties. The scientists also used a radio echo sounding (RES) imaging technique to map previous flow patterns and surface temperatures of the Greenland ice sheet to get a more complete picture of temperatures and glacial melt during the Eemian interglacial period.
            Data from the ice core showed significant spikes in the levels of CH4, d15N, d18Oatm, and N2O, and a drop in the d18Oicelevels between depths of 2,370 and 2,418 meters. These changing levels of atmospheric isotopes correspond to the time of the Eemian interglacial period, and are furthermore all associated with global temperature rise. The largest spike in CH4 occurred 128,000 years ago, which indicates that temperatures peaked around this time. Also at these depths, the air content levels in the ice increased significantly, which points to melting and refreezing events near the surface of the glacier. Using these data, the scientists calculated a surface temperature change of +7.5 ± 1.8°C at the depositional site compared to the previous millennium.
            The RES data also reveal significant folding patterns during the Eemian interglacial period, which agrees with the NEEM ice core’s data from the time period. The radar’s images show continuous and undisturbed internal layers to depths of 2,200 meters in the NEEM region. Below 2,200 meters, however, significant folding and overturning patterns due to glacial movement occurred in the ice, making it fuzzy and more difficult to interpret. The NEEM data verify these findings, as they appear to be too disturbed at depths below 2,450 meters to reconstruct temperature data. Also, the RES images show that Eemian ice contains much larger crystals than glacial ice, which is another indicator of warming and refreezing during this time period.  
            Dahl-Jenson et al. concluded that over the course of 6,000 years, between 128,000 years ago and 122,000 years ago, surface elevation decreased from 210 ± 350 above to 130 ± 300 meters below present surface elevations. When adjusted for isostatic rebound, the overall change in glacial elevation was calculated to be 400 ± 350 meters. The Eemian interglacial period saw a sea level change of 4–8 meters, and the NEEM data predict a contribution to sea level change of about two meters from the Greenland ice sheet. This indicates that Antarctica must have contributed significantly to global sea level change during this time period.
            Recent observations made in 2012 indicate strong patterns of melt at the NEEM site and parallel many changes found in the ice core from the Eemian interglacial period. Few melt layers were found before 1995 but they have increased significantly since then, and in July 2012 a large heat wave caused surface melt over 97% of the Greenland ice sheet. The 2010–12 annual surface temperatures have been recorded as 1–2°C above the 1950–80 average, and Dahl-Jenson et al. estimated a temperature change of 8 ±4°C warmer than the average of the recent millennium during the Eemian interglacial period. While this is a much greater change than what is currently being observed, similar melt patterns to those of the Eemian period have been detected and surface melt is likely to become more common in the future. 

Ice-Sheet Mass Balance Estimates Combining a Variety of Techniques

There are many different ways to measure the changes in ice mass on Earth’s polar ice sheets. Each technique has its own strengths and weaknesses, and combining the different measurements can give a more complete and accurate picture of the changes in ice-sheet mass balance. By combining the data from satellite altimetry, interferometry, and gravimetry, Shepherd et al. (2012) were able to further analyze the ice sheets in Greenland (GrIS), East Antarctica (EAIS), West Antarctica (WAIS), and the Antarctic Peninsula (APIS). They found that the changes in mass between 1992 and 2011 were      –142 ± 49 Gt yr –1 in Greenland, +14 ± 43 Gt yr –1 in East Antarctica, –65 ±26 Gt yr –1 in West Antarctica, and –20 ± 14 Gt yr –1 in the Antarctic Peninsula. This amounted to an overall change of –1350 ± 1010 Gt in the Antarctic ice sheet and –2700 ±930 Gt in that of Greenland between 1992 and 2011, equating to a sea level rise of 11.2 ± 3.8 mm during this time period.¾Olivia Jacobs
Shepherd, A., Ivins, E.R., Geruo, A., Barletta, V.R., Bentley, M.J., Bettadpur, S.,
Briggs, K.H., Bromwich, D.H., Forsberg, R., Galin, N., Horwath, M., Jacobs, S., Joughin, I., King, M.A.., Lenaerts, J.T., Li, J., Ligtenberg, S.R., Luckman, A., Luthcke, S.B., McMillan, M., Meister, R., Milne, G., Mouginot, J., Muir, A., Picolas, J.P., Paden, J., Payne, A.J., Pritchard, H., Rignot, E., Rott, H., Sorensen, L.S., Scambos, T.A., Scheuchl, B., Schrama, E.J., Smith, B., Sundal, A.V., van Angelen, J.H., van de Berg, W.J., van den Broeke, M.R., Vaughan, D.G., Velicogna, I., Wahr, J., Whitehouse, P.L., Wingham, D.J., Yi, D., Young, D., Zwally, H.J. 2012. A reconciled estimate of ice-sheet mass balance. Science 338, 1183–1189.[GSS shepherd ivins ice-sheet]
           

Shepherd et al. (2012) used a variety of data to get the most accurate picture of ice-sheet mass changes over the past 19 years. To accurately reprocess the data, they used common time intervals and definitions of the respective ice-sheets. The ice-sheet surface mass balance (SBM) estimates, collected by RACMO2 between 1979 and 2010, include solid and liquid precipitation, surface sublimation, drifting snow transport, erosion and sublimation, and meltwater formation, refreezing, retention, and runoff.
The scientists accounted for glacial isostatic adjustments (GIA)¾the changes in the ocean basins over time from unweighting of glaciers and ice sheets¾by applying six different types of satellite gravimetry and altimetry models over eight years of data collection. Radar and laser altimetry were also used to give estimates of ice-sheet mass balance through measurements of ice-sheet volume change. Radar altimetry (RA) gives precise measurements of elevation change, but is not able to accurately convert volume changes to mass changes. To do this, Shepherd et al. used laser altimetry (LA) to measure firn-layer thickness.
            The input-output method (IOM) was also applied to quantify the difference between glacier mass gained through snowfall and lost by sublimation and meltwater runoff in both Greenland and Antarctica between 1992 and 2010. This type of analysis is particularly useful because it can isolate different drainage basins and compare the different runoff volumes. Lastly, gravimetry data (including GRACE) estimate changes in ice-sheet mass balance through the ice sheet’s changing gravitational attraction. These data were particularly useful because they give regional averages and allow for monthly samplings. However, these measurements cannot account for the changes in Earth’s crust and mantle, which were then accounted for using GIA models.
            By applying these various techniques, Shepherd et al. were able to get a more accurate image of the changes in ice-sheet mass balance in Greenland and Antarctica. When RA and IOM models of 52 drainage basins in the Antarctic ice-sheet (AIS) were compared, there was an agreement of mass changes in 42 of the 52 basins. The average difference between the models was 1.4 ± 3.8 Gt yr –1, indicating that the RA and IOM models were similar. Knowing this, the scientists used IOM data to account for mass changes in the Antarctic Peninsula (APIS) where RA data were unavailable.
            Next, Shepherd et al. considered the exceptional snowfall in 2009 in East Antarctica (EAIS) to analyze whether or not geodetic techniques can detect fluctuations in SMB. They found that RACMO2, RA, and GRACE estimates, which record firn thickness and mass, volume, and mass fluctuations, respectively, were all able to detect the changes during this year of heavy snowfall.
            The scientists in this study also attempted to find a methodological intercomparison in the sampling techniques used. For this, they used the time period between October 2003 and December 2008, when all four satellite geodetic techniques (IOM, RA, LA, and gravimetry) were operating. By combining these different means of measurement, the errors in estimated ice-sheet imbalances were lessened, and the mass imbalances in the AIS and GIS were  –72 ± 43 Gt yr –1 and –232 ± 23 Gt yr –1, respectively. However, these data also revealed that the mass imbalances vary cyclically and by large amounts over 2 to 4 year periods, and thus short-term fluctuations should be acknowledged.
            The data were also combined to examine trends over both short and long time periods. This integration showed that the GrIS loss in the 1990s was modest, but increased sharply in the new millennium. Currently, there is also an increasing mass loss from WAIS and APIS. The APIS was close to being in balance in the 1990s, but loss from this ice sheet now accounts for about 25% of AIS ice-sheet loss even though it only occupies about 4% of the total area. Comparatively, the EAIS has seen a gain in mass due to an increase in snowfall, but these changes have happened over too short a time period to tell if the change is a short-term fluctuation or a long-term trend.
            In conclusion, the data compiled in this study show that between 1992 and 2011, the AIS has lost 1350 ± 1010 Gt of ice and the GrIS has lost 2700 ±930 Gt. This is equal to a sea level increase of 11.2 ± 3.8 mm. While the combinations of data in this study lessened uncertainty in the changes in ice-sheet mass balance, there is still ample space to improve current sampling techniques. The EAIS could benefit from measurements with greater spatial sampling, while the APIS could benefit from longer temporal sampling. The SMB also shows great fluctuations because of the short time period of sampling, and could thus benefit from longer spans of data. However, Shepherd et al. were able to combine existing techniques to give better estimates despite current shortcomings in the data. 

Lower satellite-gravimetry estimates of Antarctic sea-level contribution

Models of Antarctic ice loss have been extremely difficult to construct because of the shortcomings of satellite measurements. Previously, scientists have used a satellite called Gravity Recovery and Climate Experiment (GRACE), which is unable to estimate ice loss within reconcilable errors because of mass changes due to glacial isostatic adjustment (King et al. 2012). However, new imaging techniques, called W12a GIA modeling, have allowed for more accurate predictions by accounting for changes in the ocean’s basins. Consequently, new data predict that the ice-mass change in Antarctica occurs at a rate of about one third or one half the recent GRACE estimates, or about –69 ± 9 Gt yr–1. Furthermore, this new modeling suggests that ice loss in Antarctica is concentrated around the Amundsen Sea coast in West Antarctica while the ice-mass in East Antarctica is gaining substantial mass. Ultimately, the Antarctic ice sheet is likely not contributing to sea level changes as rapidly as previously predicted. —Olivia Jacobs
King, M.A., Bingham, R.J., Moore, P., Whitehouse, P.L., Bentley, M.J., Milne, G.A., 2012. Lower satellite-gravimetry estimates of Antarctic sea-level contribution. Nature 491, 586589.[GSS king milne satellite]

            King et al. (2012) used a new system of measuring ice-mass loss, which allowed for more accurate predictions and inferences than the previously used GRACE model. This is because it accounts for the changes in the glacial isostatic adjustment. Since the Last Glacial Maximum, the surface loads on the ocean floor have changed substantially, and this change in mass causes the earth’s basins to change levels. Particularly, North America is no longer covered with several kilometers of ice, so the mantle is now slowly uplifting. With new modeling techniques (W12a), the authors were able analyze 26 independent Antarctic drainage basins from August 2002 to December 2010 and account for this uplift data. 
            Across all basins, the new mass change was estimated to be –69 ± 9 Gt yr–1, indicating a significant shrinkage pattern. East Antarctica is partially compensating for the great losses in West Antarctica as it is gaining ice-mass at a rate of about +60 ±13 Gt yr­–1 while West Antarctica is losing mass at about –118 ­­± 9 Gt yr­–1. However, this rate of loss is only 36–48% the recent GRACE estimates. The authors note that the Antarctic ice sheet’s rate of loss has also been increasing over time, but not as much as previous estimates allowed. King et al. found that the rate of mass loss for the continent as a whole has been increasing over the analysis period (about nine years) by about –4 ­­± 16 Gt yr–2. This is about 15% of previous estimates, and is not statistically significant from zero.
While the earth’s sea levels have been rising at an average rate of 3.2 mm yr–1, the GRACE measurements have been too varied to give insight into which ice-masses contribute the most to this change. The W12a estimates that the Antarctic ice sheet is contributing about +0.19 ± 18 Gt yr–1 to this increase in sea levels, which is less than 10% of the total observed increase during this study’s time period. Furthermore, King et al. used similar models to predict total glacial contribution to sea levels, and estimated that they contribute about 38% of total sea level rise, while the rest is attributed to ocean warming and continental hydrology. Ultimately, future predictions suggest that the basin that contains Pine Island Glacier, a glacier along the Admundsen Sea coast, could increase sea levels by a maximum of 7.6 cm this century. Additional melt from the Antarctic ice sheet may mean that the total glacial melt could exceed this value by 2100. 

Percolation Zones in Greenland Will Hold the Ice Sheet’s Meltwater and Prevent Rises in Sea Levels for More than a Decade

Surface melt off of the Greenland ice sheet has reached record levels in 2005, 2007, 2010, and 2012. As the global temperature continues to rise, more ice will continue to melt, with severe consequences for global sea levels. However, this ice sheet’s meltwater is often refrozen in the ice sheet’s percolation zone, an area that is perennially covered by snow and partially compacted snow (firn). Here, meltwater can affect the ice sheet’s flow dynamics, sea levels, and mass balance if it melts completely and runs off, but not if it remains refozen in this percolation zone. By observing firn structure and meltwater retention of the Greenland ice sheet, Harper et al. (2012) were able to understand the flow dynamics of the meltwater and consequently predict future global impacts of the ice sheet’s shrinking size. They found that some meltwater from the Greenland ice sheet will fill pore space in the percolation zones of the ice sheet, and will thus not affect sea levels. While this pore space is not unlimited, Haper et al. suggested that water will be rerouted to these areas and will help preserve sea levels for at least 15 years.—Olivia Jacobs
Harper, J., Humphrey, N.. Pfeffer, W., Brown, J., Fettweis, X., 2012. Greeland ice-sheet
contribution to sea-level rise buffered by meltwater storage in firn. Natue 491,
240243.

            To analyze the patterns of meltwater movement, Harper et al. established fifteen study sites along the Greenland ice sheet and collected data for two years between 2007 and 2009. The team collected 34 ice cores and many firn temperatures readings and identified refrozen meltwater. Thermistor strings installed in boreholes showed thermal events under the surface of the percolation zone, and revealed reheating events in the ice sheet’s pores. These thermal data showed that refreezing events occurred well below earlier years of accumulation, which suggests that meltwater in the percolation zone moves downward through the firn instead of simply moving across the percolation zone and into the ocean. This means that pores in this area of the ice sheet act as “storage” space to hold and refreeze meltwater, thus avoiding further sea level escalation.
            After the team discovered that meltwater moved down through the pores in the percolation zone, they were able to calculate the mass of water per unit of area needed to fill remaining open pore space using ice core and radar data. The available space in different firns varied with changes in elevation because at lower elevations (below 2000 feet), where temperatures are warmer, some pore space has already been filled with ice, thus limiting its capacity to hold more meltwater. At elevations above 2000 meters, however, firn capacity is not reduced by previous meltwater infiltration and freezing. Accounting for this difference in storage with elevation, Harper et al. estimated a storage space between 322 and 1,289 gigatons throughout the entire percolation zone. 
            There are many uncertainties in this modeling, including the fluctuating rates of ice sheet melt, rainfall, firn compaction, and pore space additions from snowfall. Accounting for these uncertainties, and using the aforementioned limits of storage in the percolation zone, the most severe climate forecasting models suggest that current percolation zones will become completely filled with Greenland’s meltwater within 15 to 20 years, while more lenient models estimate that this will take about 30 years. From this data it is apparent that although the amount of meltwater from the Greenland ice sheet continues to increase as global temperatures increase, its effects on global sea levels may not become apparent for years to come. However, once the current percolation space is filled, meltwater will likely empty to the ocean and raise sea levels since porous spaces in percolation zone take decades to form. Thus current, observable rises in sea levels do not account for all dimensions of glacial melt, and sea levels may rise more rapidly in the decades ahead even if meltwater rates remain consistent. 

Percolation Zones in Greenland Will Hold the Ice Sheet’s Meltwater and Prevent Rises in Sea Levels for More than a Decade

Surface melt off of the Greenland ice sheet has reached record levels in 2005, 2007, 2010, and 2012. As the global temperature continues to rise, more ice will continue to melt, with severe consequences for global sea levels. However, this ice sheet’s meltwater is often refrozen in the ice sheet’s percolation zone, an area that is perennially covered by snow and partially compacted snow (firn). Here, meltwater can affect the ice sheet’s flow dynamics, sea levels, and mass balance if it melts completely and runs off, but not if it remains refozen in this percolation zone. By observing firn structure and meltwater retention of the Greenland ice sheet, Harper et al. (2012) were able to understand the flow dynamics of the meltwater and consequently predict future global impacts of the ice sheet’s shrinking size. They found that some meltwater from the Greenland ice sheet will fill pore space in the percolation zones of the ice sheet, and will thus not affect sea levels. While this pore space is not unlimited, Haper et al. suggested that water will be rerouted to these areas and will help preserve sea levels for at least 15 years.—Olivia Jacobs
Harper, J., Humphrey, N.. Pfeffer, W., Brown, J., Fettweis, X., 2012. Greeland ice-sheet
contribution to sea-level rise buffered by meltwater storage in firn. Nature 491,
240243.

                To analyze the patterns of meltwater movement, Harper et al. established fifteen study sites along the Greenland ice sheet and collected data for two years between 2007 and 2009. The team collected 34 ice cores and many firn temperatures readings and identified refrozen meltwater. Thermistor strings installed in boreholes showed thermal events under the surface of the percolation zone, and revealed reheating events in the ice sheet’s pores. These thermal data showed that refreezing events occurred well below earlier years of accumulation, which suggests that meltwater in the percolation zone moves downward through the firn instead of simply moving across the percolation zone and into the ocean. This means that pores in this area of the ice sheet act as “storage” space to hold and refreeze meltwater, thus avoiding further sea level escalation.
                After the team discovered that meltwater moved down through the pores in the percolation zone, they were able to calculate the mass of water per unit of area needed to fill remaining open pore space using ice core and radar data. The available space in different firns varied with changes in elevation because at lower elevations (below 2000 feet), where temperatures are warmer, some pore space has already been filled with ice, thus limiting its capacity to hold more meltwater. At elevations above 2000 meters, however, firn capacity is not reduced by previous meltwater infiltration and freezing. Accounting for this difference in storage with elevation, Harper et al. estimated a storage space between 322 and 1,289 gigatons throughout the entire percolation zone. 
                There are many uncertainties in this modeling, including the fluctuating rates of ice sheet melt, rainfall, firn compaction, and pore space additions from snowfall. Accounting for these uncertainties, and using the aforementioned limits of storage in the percolation zone, the most severe climate forecasting models suggest that current percolation zones will become completely filled with Greenland’s meltwater within 15 to 20 years, while more lenient models estimate that this will take about 30 years. From this data it is apparent that although the amount of meltwater from the Greenland ice sheet continues to increase as global temperatures increase, its effects on global sea levels may not become apparent for years to come. However, once the current percolation space is filled, meltwater will likely empty to the ocean and raise sea levels since porous spaces in percolation zone take decades to form. Thus current, observable rises in sea levels do not account for all dimensions of glacial melt, and sea levels may rise more rapidly in the decades ahead even if meltwater rates remain consistent.