Melting of Arctic Glaciers + Modelling Columbia Glacier, Alaska

---Sorry to find that this video is no longer available. I found an alternative link to the video---


I found a nice video on YouTube about melting of Arctic glaciers, and strongly recommend you to watch if you are interested in ice melting. It shows how serious the melting issue is, by looking at Columbia Glacier in Alaska and Greenland ice sheet. I really like this video, because it recorded the melting glaciers not only from scientists' view, but also from a brilliant photographer's. The video is about one and half hours, but only the first 53 minutes is worth to look at while the rest is just repeating the beginning part. The first half introduces Columbia Glacier, while the second half focuses on Greenland. 

                                                      

                                                 Created by NOVA- Geographic Special, uploaded by Mega-Documentary

I want to focus on Columbia Glacier here, because it is a nice target for modelling with abundant dataset from observations . The video documented that due to temperature increase  the glacier is melting much faster than before. It receded 10 miles over last 30 years (in the video), and the moving speed rose 8 times to 55ft per day. There is a short video showing it retreats over the past few decades. 

There is a journal article (Colgan et al., 2012) about modelling Columbia Glacier by using a 1D flowline model to simulate the form and flow of the glacier. The model was developed to tackle the transient rate of change in ice thickness via the differential equation (Equation 1, dH/dt). Variables and their units used in this study are listed below. The dominant principle used in the model was mass conservation. Based on lots of variables, relations and assumptions, the annual surface mass balance rate was parameterised as a linear function of ice surface elevation (Equation 2). Before modelling, climate variability and forcing should be carefully considered, and a decadal perturbation was introduced to simulate the climate variability and forcing. Model implementation and boundary conditions were also important. Monte Carlo ensemble filtering was applied to estimate the cumulative uncertainty from forcing and parameters, and to remove unrealistic simulations via the two-step filter. Their results (Figure 1) showed that a new dynamic equilibrium ice geometry would be reached around 2020 with a new terminus flowline position around 40 km, and the iceberg calving rate would decline back to the value before retreat. Though the simulation matched observation, there were still some limitations existing. They listed the model limitation in the article, including less diverse in time-series, effects on ice cliff height and calving flux during  ensemble  filtering, inherent limitations in the glacier density and lateral effects treatment, and loss in complexity of the real world.

Figure 1. Results from the study, adopted from Colgan et al., (2012).

In conclusion, modelling is powerful to simulate changes in the past, current and future. Like the study above, the model simulations covered the changes in Columbia Glacier from 1850 to 2100, while observations only available for a period of few decades. Future projections allow us to find out how the glacier would change in the future, which is a key contribution and benefit from modelling. However, abundant works are required before applying the model to the real world. After application, uncertainties still need to be considered. Models cannot absolutely represent the real world and predict the future, but currently we don’t have any better way to predict what would happen in the future.