Anyway... I digress...... Temperature!! important! unpredictable?? Nevertheless... with respect to glaciers.. it is hoped we can better predict their response in the future by understanding how exactly they react to variations in temperature . The important thing to note with predicting anything is that it is based on some very general assumptions and simplifications. For example knowing the exact temperature across a landscape at any one time isn't going to happen because we don't have a thermometer waving around in the air every few meters.
It is a common issue upon glaciers that we don't know how the air temperature varies. Even if we had 10 million sensors across your average glacier and all the time in the world to set them up, there are often areas that we cannot really access because they are too dangerous: either they are too high up or likely to be highly crevassed. It is typical for a glaciology study that looks at the accuracy of predictions to be operating with just a couple of detailed weather stations on a glacier.
I should mention that I've decided to produce awfully sketched up paint diagrams (or something edited in paint) to demonstrate my point(s), if indeed I have any.
Because, in the above example, the AWSs are lower on the glacier, the difference in temperature between 1 and 2 would be used to estimate the rate of temperature change further up the glacier where there are no measurements. If the AWS 2 was exactly 200 meters higher than AWS 1 and the temperature difference was 1 degree... we would assume a further 200 meters of elevation increase would be another 1 degree drop in temperature. This is just a hypothetical temperature for Storglaciaren, but that is often how temperatures are estimated.
Also, these lapse rates are often constant in time and space in many studies that predict melt. A good study on this glacier by Regine Hock and Bjorn Holmgren (2005) (1) investigates the energy balance variation by assuming a constant decrease in temperature of 5.5degrees (C) per kilometer of elevation increase (only after finding little difference in changing the lapse rates). They find that the simulations agree well with observed melting from the glacier.
However...... It has been found by many studies that assuming that temperature changes evenly and consistently as a function of elevation is often incorrect and that predicting melt from glaciers using these assumptions can be inaccurate. A fantastic study by Lene Petersen and Francesca Pellicciotti (2011) (2) shows variations in the way we can estimate temperatures across a glacier by studying the Juncal Norte Glacier in Chile. The below diagram from this study shows the differences in changing the lapse rate we use to estimate temperatures.
On top of all this difficulty, weather (such as rain and wind) can also impact the rate of decrease with elevation. For example, wind alters the way in which the free air exchanges energy with the glacier surface (known as the sensible heat flux) and this can be a major factor in determining the temperature we are interested in. But if wind is a common thing on the lower part of the glacier but not on the upper part, estimating lapse rates from the temperature impacted by localised weather is riddled with potential inaccuracies.
Welcome to my world of research!
It's very simple on the whole... but there is a lot to think about.
I have been summarising the variations in lapse rates across studies for the last decade or so and found that ultimately; the unique nature of each glacier means that assuming anything is the same can be a problem. Nevertheless... for some glaciers (i.e. Storglaciaren) changes to the lapse rate can be very negligible.
My research aims to get a really widespread set up of temperature stations across a glacier and find exactly how temperature varies while a glacier is melting in the spring/summer, what causes it to change vary and how this can affect the way we predict melting.
More to follow......
(below are the papers I've referred to... interesting and worth a read)
(1) Hock, R., & Holmgren, B. (2005). A distributed surface energy-balance model for complex, Sweden topography and its application to Storglaciaren. Journal of Glaciology, 51(172), 25–36.
(2) Petersen, L., & Pellicciotti, F. (2011). Spatial and temporal variability of air temperature on a melting glacier: Atmospheric controls, extrapolation methods and their effect on melt modeling, Juncal Norte Glacier, Chile. Journal of Geophysical Research, 116(D23), D23109. doi:10.1029/2011JD015842