Building better climate change models

Submitted by Sammy Smith (

Ranasinghe_art_title_bannerFact: Georgia has many archaeological sites along the coast at sea level or only a few feet above sea level.

Fact: Scientists have measured a global rise in sea level over the last few years.

Fact: Scientists say the sea level will continue to rise.

The question: how much will it rise?

Why do we ask this question on the Society for Georgia Archaeology’s website? Because members of the society are concerned about the impact a sea level rise will have on coastal archaeological sites. One type of coastal site is prehistoric shell mounds, that is, mounds of shells discarded by ancient diners. Other coastal archaeological sites are historic buildings like lighthouses and fishing piers and docks.

So, along with other scientists, archaeologists look to research on changes in the sea level and the impact it will have on the shoreline. This means we look to models, or scientific predictions, of how the sea level will rise, both how much and how.

Modeling such a complex situation is very difficult. A model in this sense is a carefully described if-then assessment of the factors involved, and how they interact. For example, if rainfall increases, or if temperatures increase, or if desertification increases, THEN the effect is…whatever. A robust model will incorporate many, many factors, and describe how these factors are interrelated, or how changes in one will cause changes in others.

In a recent editorial article called “Rising seas and retreating coastlines” in the professional journal Climatic Change, and available free online, Roshanka Ranasinghe and Marcel J.F. Stive discuss what we need to generate a good model for predicting climate change:

A robust solution to the problem [that is, modeling climate change] lies in comprehensive bottom–up (small-scale, process-based) and top–down (large-scale, behavior-based) numerical models. Once comprehensively validated by field data, such numerical models can be strategically applied to determine quantitative forcing-response relationships of complex, non-linear coastal processes. These relationships can then be aggregated and/or parameterized and embedded into a robust and easy-to-use numerical model which accounts for at least the primary physical processes governing coastal recession. (page 467)

There are a lot of Big Words there!

So, what do these sentences mean?

The first sentence means that a good model will take into account both local, small-scale factors (e.g., the angle of an individual island relative to offshore currents) and large-scale factors (e.g., widescale changes in landuse patterns so that vegetation cover increases).

The second sentence is a recommendation that the model be cross-checked with actual field data. In other words, it’s not enough to make a model, but a good model should be checked against data we already have to make sure they fit the model. This also makes the model more robust.

The last sentence recommends taking the factors and the existing data and incorporating them into a numerical or mathematical model that includes the factors usually discussed—like temperature, rainfall, and landuse changes—and also includes coastline processes including the impact of waves along the shore, and how sediment is transported along the shoreline.

This last is probably something you’ve not heard about with regard to the climate change debate. Still, the recommendation to include how sediments move, and how this affects landforms does seem important. After all, as the water level rises, this is the place it contacts land, and this is the place where the higher levels will change the land.

What other factors do you think are important in modeling climate change?

This website has another, older story on climate change and Georgia archaeology; find it here.