Limits...
A century of variation in the dependence of Greenland iceberg calving on ice sheet surface mass balance and regional climate change.

Bigg GR, Wei HL, Wilton DJ, Zhao Y, Billings SA, Hanna E, Kadirkamanathan V - Proc. Math. Phys. Eng. Sci. (2014)

Bottom Line: A century-long record of Greenland icebergs comes from the International Ice Patrol's record of icebergs (I48N) passing latitude 48° N, off Newfoundland.I48N exhibits strong interannual variability, with a significant increase in amplitude over recent decades.We also suggest that GrIS calving discharge is episodic on at least a regional scale and has recently been increasing significantly, largely as a result of west Greenland sources.

View Article: PubMed Central - PubMed

Affiliation: Department of Geography , University of Sheffield , Sheffield S10 2TN, UK.

ABSTRACT
Iceberg calving is a major component of the total mass balance of the Greenland ice sheet (GrIS). A century-long record of Greenland icebergs comes from the International Ice Patrol's record of icebergs (I48N) passing latitude 48° N, off Newfoundland. I48N exhibits strong interannual variability, with a significant increase in amplitude over recent decades. In this study, we show, through a combination of nonlinear system identification and coupled ocean-iceberg modelling, that I48N's variability is predominantly caused by fluctuation in GrIS calving discharge rather than open ocean iceberg melting. We also demonstrate that the episodic variation in iceberg discharge is strongly linked to a nonlinear combination of recent changes in the surface mass balance (SMB) of the GrIS and regional atmospheric and oceanic climate variability, on the scale of the previous 1-3 years, with the dominant causal mechanism shifting between glaciological (SMB) and climatic (ocean temperature) over time. We suggest that this is a change in whether glacial run-off or under-ice melting is dominant, respectively. We also suggest that GrIS calving discharge is episodic on at least a regional scale and has recently been increasing significantly, largely as a result of west Greenland sources.

No MeSH data available.


Related in: MedlinePlus

Contributions to the NARMAX model. Computed contributions of SMB, LSST and NAO to the annual iceberg numbers over 1900–2008, based on the ERR values for a 30 year sliding window, incremented 1 year at a time, where SMB, LSST and NAO were considered as inputs. The ERR value of each window for each variable was calculated by summing the ERR value of selected model terms that included the considered variable. While the model will never explain all the variance in the time series, owing to observational error and model simplifications, the sum of the contributions from SMB, LSST and NAO (always less than 1.0) is a measure of the model fit. To avoid inconsistencies due to initial and final conditions in the dataset, only those years centred in a distinct 30 year window are shown. The initial (before 1915) and final (after 1993) ERR contributions were ignored as the windows centred on these years would have been necessarily shorter than the standard 30 year window because of the finite data time series.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4042714&req=5

RSPA20130662F5: Contributions to the NARMAX model. Computed contributions of SMB, LSST and NAO to the annual iceberg numbers over 1900–2008, based on the ERR values for a 30 year sliding window, incremented 1 year at a time, where SMB, LSST and NAO were considered as inputs. The ERR value of each window for each variable was calculated by summing the ERR value of selected model terms that included the considered variable. While the model will never explain all the variance in the time series, owing to observational error and model simplifications, the sum of the contributions from SMB, LSST and NAO (always less than 1.0) is a measure of the model fit. To avoid inconsistencies due to initial and final conditions in the dataset, only those years centred in a distinct 30 year window are shown. The initial (before 1915) and final (after 1993) ERR contributions were ignored as the windows centred on these years would have been necessarily shorter than the standard 30 year window because of the finite data time series.

Mentions: The results of the full NARMAX model of I48N are shown in figure 2 and demonstrate a high level of fit (r=0.84; see table 1 for characteristic ERR values, but it should be remembered that these are just two realizations of models from the 79 sliding window intervals). Figure 5 shows the evolving contributions of SMB, NAO and LSST to the variance explained by the NARMAX model. For the first half of the twentieth century, the contribution of SMB to the model, based on summing the ERR values, is strongly dominant, explaining 50–60% of the variance in I48N. There is then a period when the climatic indices of the NAO, and particularly LSST, explain similar, or greater, amounts of the variation in I48N compared with the SMB. However, in recent years, SMB has reverted to being dominant. This late twentieth century importance of the climate-related indices is not directly linked to climate change affecting iceberg trajectories, as the ocean model (figure 2) shows little change in the mean number of icebergs reaching 48° N over the last few decades, even though the NAO and LSST have generally been higher (figure 1).Table 1.


A century of variation in the dependence of Greenland iceberg calving on ice sheet surface mass balance and regional climate change.

Bigg GR, Wei HL, Wilton DJ, Zhao Y, Billings SA, Hanna E, Kadirkamanathan V - Proc. Math. Phys. Eng. Sci. (2014)

Contributions to the NARMAX model. Computed contributions of SMB, LSST and NAO to the annual iceberg numbers over 1900–2008, based on the ERR values for a 30 year sliding window, incremented 1 year at a time, where SMB, LSST and NAO were considered as inputs. The ERR value of each window for each variable was calculated by summing the ERR value of selected model terms that included the considered variable. While the model will never explain all the variance in the time series, owing to observational error and model simplifications, the sum of the contributions from SMB, LSST and NAO (always less than 1.0) is a measure of the model fit. To avoid inconsistencies due to initial and final conditions in the dataset, only those years centred in a distinct 30 year window are shown. The initial (before 1915) and final (after 1993) ERR contributions were ignored as the windows centred on these years would have been necessarily shorter than the standard 30 year window because of the finite data time series.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4042714&req=5

RSPA20130662F5: Contributions to the NARMAX model. Computed contributions of SMB, LSST and NAO to the annual iceberg numbers over 1900–2008, based on the ERR values for a 30 year sliding window, incremented 1 year at a time, where SMB, LSST and NAO were considered as inputs. The ERR value of each window for each variable was calculated by summing the ERR value of selected model terms that included the considered variable. While the model will never explain all the variance in the time series, owing to observational error and model simplifications, the sum of the contributions from SMB, LSST and NAO (always less than 1.0) is a measure of the model fit. To avoid inconsistencies due to initial and final conditions in the dataset, only those years centred in a distinct 30 year window are shown. The initial (before 1915) and final (after 1993) ERR contributions were ignored as the windows centred on these years would have been necessarily shorter than the standard 30 year window because of the finite data time series.
Mentions: The results of the full NARMAX model of I48N are shown in figure 2 and demonstrate a high level of fit (r=0.84; see table 1 for characteristic ERR values, but it should be remembered that these are just two realizations of models from the 79 sliding window intervals). Figure 5 shows the evolving contributions of SMB, NAO and LSST to the variance explained by the NARMAX model. For the first half of the twentieth century, the contribution of SMB to the model, based on summing the ERR values, is strongly dominant, explaining 50–60% of the variance in I48N. There is then a period when the climatic indices of the NAO, and particularly LSST, explain similar, or greater, amounts of the variation in I48N compared with the SMB. However, in recent years, SMB has reverted to being dominant. This late twentieth century importance of the climate-related indices is not directly linked to climate change affecting iceberg trajectories, as the ocean model (figure 2) shows little change in the mean number of icebergs reaching 48° N over the last few decades, even though the NAO and LSST have generally been higher (figure 1).Table 1.

Bottom Line: A century-long record of Greenland icebergs comes from the International Ice Patrol's record of icebergs (I48N) passing latitude 48° N, off Newfoundland.I48N exhibits strong interannual variability, with a significant increase in amplitude over recent decades.We also suggest that GrIS calving discharge is episodic on at least a regional scale and has recently been increasing significantly, largely as a result of west Greenland sources.

View Article: PubMed Central - PubMed

Affiliation: Department of Geography , University of Sheffield , Sheffield S10 2TN, UK.

ABSTRACT
Iceberg calving is a major component of the total mass balance of the Greenland ice sheet (GrIS). A century-long record of Greenland icebergs comes from the International Ice Patrol's record of icebergs (I48N) passing latitude 48° N, off Newfoundland. I48N exhibits strong interannual variability, with a significant increase in amplitude over recent decades. In this study, we show, through a combination of nonlinear system identification and coupled ocean-iceberg modelling, that I48N's variability is predominantly caused by fluctuation in GrIS calving discharge rather than open ocean iceberg melting. We also demonstrate that the episodic variation in iceberg discharge is strongly linked to a nonlinear combination of recent changes in the surface mass balance (SMB) of the GrIS and regional atmospheric and oceanic climate variability, on the scale of the previous 1-3 years, with the dominant causal mechanism shifting between glaciological (SMB) and climatic (ocean temperature) over time. We suggest that this is a change in whether glacial run-off or under-ice melting is dominant, respectively. We also suggest that GrIS calving discharge is episodic on at least a regional scale and has recently been increasing significantly, largely as a result of west Greenland sources.

No MeSH data available.


Related in: MedlinePlus