cc: ivjansse@vub.ac.be date: Wed, 14 Jul 2004 11:24:32 +0100 from: Sarah Raper subject: Re: initialisation problems to: Philippe Huybrechts OK, we will look into this next week. I would have liked Ives to send me some graphs but I didn't get any. > This went up to a point > that Ives thought the 1961-1990 rate of glacier > volume change was a prescribed quantity, which > sounded rather absurd to me This is the case. Rate of change (which is mass balance) depends on ELA. Can fine tune ELA to give sensible rate of change over 1961-1990. Glaciers were not in equilibrium 1961-1990, I think. On 13 Jul 2004, at 13:12, Philippe Huybrechts wrote: > Sarah: > > we've done all these things, even starting from the > year 1000, with and without climate change. The most > robust feature of all these runs is that the rate of > glacier retreat for the second half of the 20th > century is always the same, irrespective of the > starting point of the climate forcing, and > irrespective of whether the 20th century climate was > cooling, warming, or stable. This went up to a point > that Ives thought the 1961-1990 rate of glacier > volume change was a prescribed quantity, which > sounded rather absurd to me. But I thought Ives > contacted/ spoke to you last week or the week before > to explain all this. Now Ives has decided not to show > up for the next 5 weeks and I don't have all his > material with me. > > BTW, you didn't answer my question: does your > algorithm work in anomaly model or not?? > > Please seriously consider the points raised above as > well and tell me about your results from your basic > testing of your algorithm. > > We can also talk about it next week at AWI: please > convince me your model is doing what it should do in > the sense that glacier growth and retreat somehow is > in accordance with a cooling or a warming, > respectively. We might also have implemented your > algorithm wrongly, of course, as Ives had to > intervene in the code at quite a number of places to > avoid all kinds of computer crashes. > > Philippe. > > > > ---------------------------------------------------- > | Dr. Philippe Huybrechts > | Alfred-Wegener-Institut fuer Polar- und > | Meeresforschung > | Postfach 120161 > | D-27515 Bremerhaven > | tel: +49-471-4831-1194 > | fax: +49-471-4831-1149 > | E-mail: phuybrechts@awi-bremerhaven.de > | www.awi-bremerhaven.de/People/show?uid=phuybrechts > ---------------------------------------------------- > > > ----- Original Message ----- > From: Sarah Raper > Date: Tuesday, July 13, 2004 11:22 am > Subject: Re: initialisation problems > >> It took a while to find a suitable reply to your email. Try looking >> at >> my Storlaciaren >> paper, otherwise I suggest you try the following. >> 1) Does the model 'work' when starting with your anomalies from 1900? >> 2) What about when you start from 1850? >> 3) 1800 etc. >> Find when and where (region) it fails and print out the anomalies >> for >> that region. >> >> On 11 Jul 2004, at 20:09, Philippe Huybrechts wrote: >> >>> Dear Ives and Sarah, >>> >>> I don't quite understand what is going on here. I >>> thought the glacier model relied on temperature >>> anomalies which were independently calculated from >>> any forcing time series by subtracting them from >>> their own 1961-1990 mean? In that case, it does not >>> matter which climate model is used, does it? Relying >>> on absolute climate forcing, on the other hand, is >>> certainly going to introduce a lot of systematic >>> biases (climatic as well as topographic, the latter >>> very important since in ECBILT/CLIO the surface >>> topography is heavily smoothed and the glacier >>> algorithm does not know how). >>> >>> Sarah please explain this in detail to me as I am >>> starting to get more questions than answers from your >>> algorithm. It would also be good to see the results >>> of your basic testing of your code. Surely when the >>> model is driven by a constan 1961-1990 climate I >>> would like to see the glaciers being in steady state. >>> Anything else gives me an uneasy feeling about it. >>> Looking at Morteratsch (which you already did, did >>> you?) may well distract from teh main issue, at least >>> in my eyes.... >>> >>> Best wishes, >>> >>> Philippe. >>> >>> ---------------------------------------------------- >>> | Dr. Philippe Huybrechts >>> | Alfred-Wegener-Institut fuer Polar- und >>> | Meeresforschung >>> | Postfach 120161 >>> | D-27515 Bremerhaven >>> | tel: +49-471-4831-1194 >>> | fax: +49-471-4831-1149 >>> | E-mail: phuybrechts@awi-bremerhaven.de >>> | www.awi-bremerhaven.de/People/show?uid=phuybrechts >>> ---------------------------------------------------- >>> >>> >>> ----- Original Message ----- >>> From: sraper@awi-bremerhaven.de >>> Date: Thursday, July 1, 2004 5:43 pm >>> Subject: Re: initialisation problems >>> >>>> Dear Ives, >>>> >>>> I have been thinking about this a bit and though changing the ELA >>>> is an option there is another question to answer first, I think. >>>> The >>>> question is whether the climate that ECBILT/CLIO (or whatever its >>>> called) produces is compatable with the present observed area of >>>> glaciers. For this is what we assume by trying to get glacier melt >>>> over the period with correct present glacier area. >>>> >>>> I suggest we try to test this using just one glaicer. Lets do >>>> Morteratsch. If you can extract for me the temperature input for >>>> just this one grid square and send it to me I will try it out. Then >>>> we >>>> will see..... >>>> >>>> Regards, >>>> >>>> Sarah >>>> >>>> >>> >> >> >