date: Thu, 15 Jul 2004 11:29:22 +0100 from: Sarah Raper subject: Re: tanx to: Tom Wigley Dear Tom, find the **'s look also a end. I do this because my emails have a habit of disappearing as I write them so I use an editor. Sarah, Thanx for the comments. I will revise accordingly. Re 1880, you can fill me in on this. In the TAR there is a gs(1990) term, I presume is is the area-corrected melt from 1880 to 1990. There is also a gu(1990). In Table 11.17 there are model by model values of g(1990), which the text on p. 682 implies is gu(1990). The gu/gs difference in 1990 based on the TAR formula is very small. Both of these are about 2cm. Just what period this covers is not clear. The p. 682 text doesn't say, but it does say that T(1990) is from 'the late 19th century'. Elsewhere, Table 11.10 says that the average GSIC rate of increase over 1910-1990 is 0.3 +/- 0.1 mm/yr. This gives a GSIC sea level rise of 2.4 +/- 0.8 cm over 1910-1990. For the same rate over 1880-1990 the rise would be 3.3 +/- 1.1 cm. So a good and consistent guess for 1880-1990 is 3 +/- 1 cm. This is probably inconsistent with Table 11.17. I have never worried about this because the pre 1990 changes that can be computed from the formulae are incomplete and uncertain. I will have to be careful about this. Anyhow, I am certainly not extending the TAR formula back before 1990. Everything begins at 1990, where I take t=0. In addition, there is a g(1990) term that I simply add in order to get the full melt from (nominally) 1880. The gs/gu thing complicates the algebra a bit, but what I do is very simple.Bahr, 1996; Bahr, 1997). ** p650 last para 11.2.2 Using these methods,Van de Wal and Wild (2001) found that contraction of area reduces the estimated glacier net mass loss over the next 70 years by 15 to 20% (see also Section 11.5.1.1). ** This is all I can find. It implies to me that for AOGCM glacier melt predictions the correction was applied. It simply does not mention (that I can find) what he did pre 1990. So best just avoid the issue as you seem to suggest. Your V0 of around 30cm is good. However, I would like to stick to the TAR estimate, which is 40cm as explained in the text, for this paper. **We can go with 40cm if you want, but think about the following. From the figure I sent you icecaps account for 30% of the area (12/40*100). However, they account for 125% of the volume (.15/.12*100). That's because many many glaciers are very small and shallow whereas the icecaps are much larger and therefore deeper (see histograms). For around Antarctica and Greenland there is a larger proportion of icecaps to glaciers relative to the globe. It doesn't bother me too much at least 30cm is just within the range but I'm pretty sure it will turn out to be a better estimate. Tom. ****PS Thinking about it in general: I think we should go ahead but I have still a slight reservation because it does in fact make the compliment of using his method which I dislike very much. Physically it is all wrong. Never mind, though what I did do at some point and we could do it is tune our original model to give similar results, I found the shape was a bit different. I wonder how it would compare for the longer times. Its rather easy to do espically as we know the Vo to use (the same). *** I think we need to include the Gregory and Oerlemans ref: Gregory, J.M. and J. Oerlemans, 1998: Simulated future sea level rise due to glacier melt based on regionally and seasonally resolved temperature changes. Nature, 391, 474-6. This is the official ref for the 'method', prior to the area correction. also should include: Van de Wal, R.S.W. and M. Wild, 2001: Modelling the response of glaciers to climate change, applying volume area scaling in combination with a high-resolution GCM. IMAUReport R-01-06, Utrecht University, Netherlands. there must be a better reference than this now.... On 14 Jul 2004, at 17:30, Tom Wigley wrote: > Sarah, > > Thanx for the comments. I will revise accordingly. Re 1880, you can > fill me in on this. In the TAR there is a gs(1990) term, I presume is > is the area-corrected melt from 1880 to 1990. There is also a gu(1990). > In Table 11.17 there are model by model values of g(1990), which the > text on p. 682 implies is gu(1990). The gu/gs difference in 1990 based > on the TAR formula is very small. Both of these are about 2cm. > > Just what period this covers is not clear. The p. 682 text doesn't say, > but it does say that T(1990) is from 'the late 19th century'. > > Elsewhere, Table 11.10 says that the average GSIC rate of increase over > 1910-1990 is 0.3 +/- 0.1 mm/yr. This gives a GSIC sea level rise of 2.4 > +/- 0.8 cm over 1910-1990. For the same rate over 1880-1990 the rise > would be 3.3 +/- 1.1 cm. So a good and consistent guess for 1880-1990 > is 3 +/- 1 cm. This is probably inconsistent with Table 11.17. > > I have never worried about this because the pre 1990 changes that can > be computed from the formulae are incomplete and uncertain. > > I will have to be careful about this. Anyhow, I am certainly not > extending > the TAR formula back before 1990. Everything begins at 1990, where I > take t=0. In addition, there is a g(1990) term that I simply add in > order to > get the full melt from (nominally) 1880. The gs/gu thing complicates > the > algebra a bit, but what I do is very simple. > > Your V0 of around 30cm is good. However, I would like to stick to the > TAR estimate, which is 40cm as explained in the text, for this paper. > > Tom. >