From schimel at ucar.edu Fri Sep 1 16:40:23 2006 From: schimel at ucar.edu (David Schimel) Date: Sat Sep 2 15:43:13 2006 Subject: [CCSM-Biogeochemistry] Re: [CCSM-LandWG] Physical component biases and BGC: Koppen climate-vegetation classification as a metric In-Reply-To: <44F730FB.1000008@ucar.edu> References: <44F730FB.1000008@ucar.edu> Message-ID: <44F8B6D7.7090400@ucar.edu> we are going to evaluate a third generation climate model against 1930s biogeography? What a step forward! David Lawrence wrote: > Hi everyone, > > At the Aspen Global Change Institute meeting on ESMs a few weeks ago > there was discussion of metrics for this new class of model. A metric > that GFDL is using (or intends to use) is the Koppen > climate-vegetation classification. The Koppen climate classification > characterizes 11 distinct seasonal temperature and precipitation > regimes and their associated vegetation classes (see classification > rules below). It is an efficient means of identifying climate biases > that are likely to affect carbon cycle modeling. > > I coded up the classification procedure (note that I am using the > modified Koppen classification by Guetter and Kutzback, 1990) and ran > it through some CCSM data as a first check to see whether or not it > could be useful in terms of directing model development efforts. > > I attach two plots that were created using CCSM3 data (T85 fully > coupled and AMIP, and T42 AMIP). The percent land area mis-classified > is around 41% for T85 CCSM3, 43% for T85 CAM3, and 37% for T42 CAM3. > By qualitatively examining the plots you can clearly identify, for > example, two known problems with CCSM climate. The first is that the > model is dry over the Amazon (and wet over tropical Africa) and the > Koppen classification indicates that the simulated climate will not > support tropical evergreen trees there. The second is that tundra > extends too far south, due to the cold summer bias in high latitudes > (possibly related to low simulated NPP at high latitudes?). > > Anyway, I am proposing this as a potential additional metric to be > provided to the AMWG. It can be used either quantitatively or > qualitatively and is an efficient means of identifying issues with the > simulated climate that are directly relevant to carbon cycle modeling. > > Regards, > > Dave > > Koppen classification rules: > > A Tropical rainy climate in which the temperature of the coolest > month is greater than 18C > Af Rainfall in driest month exceeds 6cm. Typical veg: evergreen > tropical rainforest > Aw Rainfall in driest month is less than 6cm. Typical veg: > tropical deciduous forests and savannas > > B Dry climates. These climates occur where R, the annual > rainfall in cm is less than the amounts given by the following formulas > R = 2T if the rainfall max occurs in winter (the wettest > winter month has more than 3x rainfall as driest summer month) > R = 2T + 7 if rainfall is evenly distributed throughout the year > R = 2T + 14 if rainfall max occurs in summer (the wettest summer > month has more than 10x the rainfall as the driest winter month) > BS Semiarid climate. Rainfall exceeds one-half the amount given by > the formulas for teh boundary between dry and hmid climatees. Typical > veg: grasslands, dry tropical savannas, low shrubs > BW Arid climates. Rainfall is less than one half the amount given > by the formulas for the boundary between dry and humid climates. > Typical veg: very sparse or lacking > > C Warm temperate rainy climates. The temp of the coolest month > is between 18 and -3 C > Cs Summer drought. Rainfall of wettest winter month is at least > 3x that of driest summer month. Typical veg: trees adapted to summer > dryness such as oaks, ecualyptus or olives > Cw Winter drought. Rainfall of wettest summer month is 10x that > of driesst winter month. > Cf No drought period. Typical veg: deciduous forests, subtropical > evergreen forests. > > D Boreal climate. Temp of the coolest month is less than -3C and > temp of warmest month is greater than 10C. Typical veg: Evergreen > coniferous forest, mixed deciduous and evergreen forests near southern > boundary > Dw Winter drought. Rainfall of wettest summer month is at least > 10x that of driest winter month > Df No drought period. > > E Snow climates. Temp of warmest month is less than 10C > ET Tundra. Temp of warmest month is greater than 0C. Typical veg: > Tundra, no trees > EF Perpetual frost. Temp of warmest month is less than 0C > > Order of classification E, B, A, C, D. For B, C, D climates summer is > warmest 3 months of year, winter is coldest 3 months of year. > > > ------------------------------------------------------------------------ > > > ------------------------------------------------------------------------ > > ------------------------------------------------------------------------ > > _______________________________________________ > CCSM-Land mailing list > CCSM-Land@cgd.ucar.edu > http://www.cgd.ucar.edu/mailman/listinfo/ccsm-land > From pjr at ucar.edu Sun Sep 3 11:51:06 2006 From: pjr at ucar.edu (Phil Rasch) Date: Sun Sep 3 13:15:32 2006 Subject: [CCSM-Biogeochemistry] Re: [CCSM-LandWG] Physical component biases and BGC: Koppen climate-vegetation classification as a metric In-Reply-To: <44F8B6D7.7090400@ucar.edu> References: <44F730FB.1000008@ucar.edu> <44F8B6D7.7090400@ucar.edu> Message-ID: <20060903175106.GA3272@ucar.edu> Its a step forward if we have been building models without comparing to something like this before! Phil On Fri, Sep 01, 2006 at 04:40:23PM -0600, David Schimel wrote: > we are going to evaluate a third generation climate model against 1930s > biogeography? What a step forward! > > David Lawrence wrote: > >Hi everyone, > > > >At the Aspen Global Change Institute meeting on ESMs a few weeks ago > >there was discussion of metrics for this new class of model. A metric > >that GFDL is using (or intends to use) is the Koppen > >climate-vegetation classification. The Koppen climate classification > >characterizes 11 distinct seasonal temperature and precipitation > >regimes and their associated vegetation classes (see classification > >rules below). It is an efficient means of identifying climate biases > >that are likely to affect carbon cycle modeling. > > > >I coded up the classification procedure (note that I am using the > >modified Koppen classification by Guetter and Kutzback, 1990) and ran > >it through some CCSM data as a first check to see whether or not it > >could be useful in terms of directing model development efforts. > > > >I attach two plots that were created using CCSM3 data (T85 fully > >coupled and AMIP, and T42 AMIP). The percent land area mis-classified > >is around 41% for T85 CCSM3, 43% for T85 CAM3, and 37% for T42 CAM3. > >By qualitatively examining the plots you can clearly identify, for > >example, two known problems with CCSM climate. The first is that the > >model is dry over the Amazon (and wet over tropical Africa) and the > >Koppen classification indicates that the simulated climate will not > >support tropical evergreen trees there. The second is that tundra > >extends too far south, due to the cold summer bias in high latitudes > >(possibly related to low simulated NPP at high latitudes?). > > > >Anyway, I am proposing this as a potential additional metric to be > >provided to the AMWG. It can be used either quantitatively or > >qualitatively and is an efficient means of identifying issues with the > >simulated climate that are directly relevant to carbon cycle modeling. > > > >Regards, > > > >Dave > > > >Koppen classification rules: > > > >A Tropical rainy climate in which the temperature of the coolest > >month is greater than 18C > >Af Rainfall in driest month exceeds 6cm. Typical veg: evergreen > >tropical rainforest > >Aw Rainfall in driest month is less than 6cm. Typical veg: > >tropical deciduous forests and savannas > > > >B Dry climates. These climates occur where R, the annual > >rainfall in cm is less than the amounts given by the following formulas > > R = 2T if the rainfall max occurs in winter (the wettest > >winter month has more than 3x rainfall as driest summer month) > > R = 2T + 7 if rainfall is evenly distributed throughout the year > > R = 2T + 14 if rainfall max occurs in summer (the wettest summer > >month has more than 10x the rainfall as the driest winter month) > >BS Semiarid climate. Rainfall exceeds one-half the amount given by > >the formulas for teh boundary between dry and hmid climatees. Typical > >veg: grasslands, dry tropical savannas, low shrubs > >BW Arid climates. Rainfall is less than one half the amount given > >by the formulas for the boundary between dry and humid climates. > >Typical veg: very sparse or lacking > > > >C Warm temperate rainy climates. The temp of the coolest month > >is between 18 and -3 C > >Cs Summer drought. Rainfall of wettest winter month is at least > >3x that of driest summer month. Typical veg: trees adapted to summer > >dryness such as oaks, ecualyptus or olives > >Cw Winter drought. Rainfall of wettest summer month is 10x that > >of driesst winter month. > >Cf No drought period. Typical veg: deciduous forests, subtropical > >evergreen forests. > > > >D Boreal climate. Temp of the coolest month is less than -3C and > >temp of warmest month is greater than 10C. Typical veg: Evergreen > >coniferous forest, mixed deciduous and evergreen forests near southern > >boundary > >Dw Winter drought. Rainfall of wettest summer month is at least > >10x that of driest winter month > >Df No drought period. > > > >E Snow climates. Temp of warmest month is less than 10C > >ET Tundra. Temp of warmest month is greater than 0C. Typical veg: > >Tundra, no trees > >EF Perpetual frost. Temp of warmest month is less than 0C > > > >Order of classification E, B, A, C, D. For B, C, D climates summer is > >warmest 3 months of year, winter is coldest 3 months of year. > > > > > >------------------------------------------------------------------------ > > > > > >------------------------------------------------------------------------ > > > >------------------------------------------------------------------------ > > > >_______________________________________________ > >CCSM-Land mailing list > >CCSM-Land@cgd.ucar.edu > >http://www.cgd.ucar.edu/mailman/listinfo/ccsm-land > > > _______________________________________________ > CCSM-Land mailing list > CCSM-Land@cgd.ucar.edu > http://www.cgd.ucar.edu/mailman/listinfo/ccsm-land From kingaw at ornl.gov Tue Sep 5 09:02:39 2006 From: kingaw at ornl.gov (King, Anthony Wayne) Date: Tue Sep 5 09:15:58 2006 Subject: [CCSM-Biogeochemistry] Re: [CCSM-LandWG] Physical component biases and BGC: Koppen climate-vegetation classification as a metric In-Reply-To: <20060903175106.GA3272@ucar.edu> Message-ID: Has anyone "overlaid" a recent observational climatology with satellite derived vegetation classification and evaluated how the "coincidence" matches the Koppen classification rules? Might be the basis of an updated Koppen-style climate classification/biogeography. Tony > -----Original Message----- > From: ccsm-biogeochemistry-bounces@cgd.ucar.edu [mailto:ccsm- > biogeochemistry-bounces@cgd.ucar.edu] On Behalf Of Phil Rasch > Sent: Sunday, September 03, 2006 1:51 PM > To: David Schimel > Cc: David Lawrence; ccsm-biogeochemistry@cgd.ucar.edu; ccsm- > land@cgd.ucar.edu > Subject: [CCSM-Biogeochemistry] Re: [CCSM-LandWG] Physical component > biases and BGC: Koppen climate-vegetation classification as a metric > > Its a step forward if we have been building models without comparing > to something like this before! > > Phil > > On Fri, Sep 01, 2006 at 04:40:23PM -0600, David Schimel wrote: > > we are going to evaluate a third generation climate model against 1930s > > biogeography? What a step forward! > > > > David Lawrence wrote: > > >Hi everyone, > > > > > >At the Aspen Global Change Institute meeting on ESMs a few weeks ago > > >there was discussion of metrics for this new class of model. A metric > > >that GFDL is using (or intends to use) is the Koppen > > >climate-vegetation classification. The Koppen climate classification > > >characterizes 11 distinct seasonal temperature and precipitation > > >regimes and their associated vegetation classes (see classification > > >rules below). It is an efficient means of identifying climate biases > > >that are likely to affect carbon cycle modeling. > > > > > >I coded up the classification procedure (note that I am using the > > >modified Koppen classification by Guetter and Kutzback, 1990) and ran > > >it through some CCSM data as a first check to see whether or not it > > >could be useful in terms of directing model development efforts. > > > > > >I attach two plots that were created using CCSM3 data (T85 fully > > >coupled and AMIP, and T42 AMIP). The percent land area mis-classified > > >is around 41% for T85 CCSM3, 43% for T85 CAM3, and 37% for T42 CAM3. > > >By qualitatively examining the plots you can clearly identify, for > > >example, two known problems with CCSM climate. The first is that the > > >model is dry over the Amazon (and wet over tropical Africa) and the > > >Koppen classification indicates that the simulated climate will not > > >support tropical evergreen trees there. The second is that tundra > > >extends too far south, due to the cold summer bias in high latitudes > > >(possibly related to low simulated NPP at high latitudes?). > > > > > >Anyway, I am proposing this as a potential additional metric to be > > >provided to the AMWG. It can be used either quantitatively or > > >qualitatively and is an efficient means of identifying issues with the > > >simulated climate that are directly relevant to carbon cycle modeling. > > > > > >Regards, > > > > > >Dave > > > > > >Koppen classification rules: > > > > > >A Tropical rainy climate in which the temperature of the coolest > > >month is greater than 18C > > >Af Rainfall in driest month exceeds 6cm. Typical veg: evergreen > > >tropical rainforest > > >Aw Rainfall in driest month is less than 6cm. Typical veg: > > >tropical deciduous forests and savannas > > > > > >B Dry climates. These climates occur where R, the annual > > >rainfall in cm is less than the amounts given by the following formulas > > > R = 2T if the rainfall max occurs in winter (the wettest > > >winter month has more than 3x rainfall as driest summer month) > > > R = 2T + 7 if rainfall is evenly distributed throughout the year > > > R = 2T + 14 if rainfall max occurs in summer (the wettest summer > > >month has more than 10x the rainfall as the driest winter month) > > >BS Semiarid climate. Rainfall exceeds one-half the amount given > by > > >the formulas for teh boundary between dry and hmid climatees. Typical > > >veg: grasslands, dry tropical savannas, low shrubs > > >BW Arid climates. Rainfall is less than one half the amount > given > > >by the formulas for the boundary between dry and humid climates. > > >Typical veg: very sparse or lacking > > > > > >C Warm temperate rainy climates. The temp of the coolest month > > >is between 18 and -3 C > > >Cs Summer drought. Rainfall of wettest winter month is at least > > >3x that of driest summer month. Typical veg: trees adapted to summer > > >dryness such as oaks, ecualyptus or olives > > >Cw Winter drought. Rainfall of wettest summer month is 10x that > > >of driesst winter month. > > >Cf No drought period. Typical veg: deciduous forests, > subtropical > > >evergreen forests. > > > > > >D Boreal climate. Temp of the coolest month is less than -3C > and > > >temp of warmest month is greater than 10C. Typical veg: Evergreen > > >coniferous forest, mixed deciduous and evergreen forests near southern > > >boundary > > >Dw Winter drought. Rainfall of wettest summer month is at least > > >10x that of driest winter month > > >Df No drought period. > > > > > >E Snow climates. Temp of warmest month is less than 10C > > >ET Tundra. Temp of warmest month is greater than 0C. Typical > veg: > > >Tundra, no trees > > >EF Perpetual frost. Temp of warmest month is less than 0C > > > > > >Order of classification E, B, A, C, D. For B, C, D climates summer is > > >warmest 3 months of year, winter is coldest 3 months of year. > > > > > > > > >----------------------------------------------------------------------- > - > > > > > > > > >----------------------------------------------------------------------- > - > > > > > >----------------------------------------------------------------------- > - > > > > > >_______________________________________________ > > >CCSM-Land mailing list > > >CCSM-Land@cgd.ucar.edu > > >http://www.cgd.ucar.edu/mailman/listinfo/ccsm-land > > > > > _______________________________________________ > > CCSM-Land mailing list > > CCSM-Land@cgd.ucar.edu > > http://www.cgd.ucar.edu/mailman/listinfo/ccsm-land > _______________________________________________ > CCSM-Biogeochemistry mailing list > CCSM-Biogeochemistry@cgd.ucar.edu > http://www.cgd.ucar.edu/mailman/listinfo/ccsm-biogeochemistry From swr at ntsg.umt.edu Tue Sep 5 10:51:50 2006 From: swr at ntsg.umt.edu (Steve Running) Date: Tue Sep 5 11:15:50 2006 Subject: [CCSM-Biogeochemistry] Re: [CCSM-LandWG] Physical component biasesand BGC: Koppen climate-vegetation classification as a metric In-Reply-To: Message-ID: <001101c6d10b$959ae680$c662080a@ntsg.umt.edu> Or try our Growing Season Index from Jolly et al 2005 that integrates the climatology. See attached. This really is a foundation for a climatology defined vegetation classification ------------------------------------------ Professor Steven W. Running, Director Numerical Terradynamic Simulation Group Dept. of Ecosystem and Conservation Sciences University of Montana Missoula, MT 59812 USA Homepage: http://www.ntsg.umt.edu/ Email: swr@ntsg.umt.edu Phone: 1-406-243-6311 FAX: 1-406-243-4510 ------------------------------------------- -------------- next part -------------- A non-text attachment was scrubbed... Name: Jolly-et-al-GCB-11-619-632.pdf Type: application/pdf Size: 738711 bytes Desc: not available Url : http://www.cgd.ucar.edu/pipermail/ccsm-biogeochemistry/attachments/20060905/610e9103/Jolly-et-al-GCB-11-619-632-0001.pdf From esundqui at usgs.gov Tue Sep 5 11:31:16 2006 From: esundqui at usgs.gov (Eric T. Sundquist) Date: Tue Sep 5 11:35:22 2006 Subject: [CCSM-Biogeochemistry] Re: [CCSM-LandWG] Physical component biases and BGC: Koppen climate-vegetation classification as a metric In-Reply-To: References: Message-ID: <44FDB464.2020307@usgs.gov> I agree with Dave S's comment that revisiting climate-based biogeographic zones would not be productive. However, I like Dave L's careful choice of words: "It is an efficient means of identifying climate biases that are likely to affect carbon cycle modeling." Basically the Koppen classes can be viewed as climate classes that are significant to vegetation and hence C cycle models. There is certainly a need for quick vegetation-relevant metrics that can be applied to the physical climate model. Simple land-area mis-classificiation metrics treat all mismatches as equally important. Perhaps a next step would be to evaluate the extent to which mismatches would be critical for climate/veg/carbon feedbacks -- for example, those that might have a lot of leverage on biomass, ET, and/or albedo. It wouldn't be difficult to quantify these. This approach would probably require adding PET or an aridity index (a la Holdridge -- I can hear Dave S groaning now!). Eric King, Anthony Wayne wrote: >Has anyone "overlaid" a recent observational climatology with satellite >derived vegetation classification and evaluated how the "coincidence" >matches the Koppen classification rules? Might be the basis of an >updated Koppen-style climate classification/biogeography. >Tony > > > >>-----Original Message----- >>From: ccsm-biogeochemistry-bounces@cgd.ucar.edu [mailto:ccsm- >>biogeochemistry-bounces@cgd.ucar.edu] On Behalf Of Phil Rasch >>Sent: Sunday, September 03, 2006 1:51 PM >>To: David Schimel >>Cc: David Lawrence; ccsm-biogeochemistry@cgd.ucar.edu; ccsm- >>land@cgd.ucar.edu >>Subject: [CCSM-Biogeochemistry] Re: [CCSM-LandWG] Physical component >>biases and BGC: Koppen climate-vegetation classification as a metric >> >>Its a step forward if we have been building models without comparing >>to something like this before! >> >>Phil >> >>On Fri, Sep 01, 2006 at 04:40:23PM -0600, David Schimel wrote: >> >> >>>we are going to evaluate a third generation climate model against >>> >>> >1930s > > >>>biogeography? What a step forward! >>> >>>David Lawrence wrote: >>> >>> >>>>Hi everyone, >>>> >>>>At the Aspen Global Change Institute meeting on ESMs a few weeks >>>> >>>> >ago > > >>>>there was discussion of metrics for this new class of model. A >>>> >>>> >metric > > >>>>that GFDL is using (or intends to use) is the Koppen >>>>climate-vegetation classification. The Koppen climate >>>> >>>> >classification > > >>>>characterizes 11 distinct seasonal temperature and precipitation >>>>regimes and their associated vegetation classes (see classification >>>>rules below). It is an efficient means of identifying climate >>>> >>>> >biases > > >>>>that are likely to affect carbon cycle modeling. >>>> >>>>I coded up the classification procedure (note that I am using the >>>>modified Koppen classification by Guetter and Kutzback, 1990) and >>>> >>>> >ran > > >>>>it through some CCSM data as a first check to see whether or not it >>>>could be useful in terms of directing model development efforts. >>>> >>>>I attach two plots that were created using CCSM3 data (T85 fully >>>>coupled and AMIP, and T42 AMIP). The percent land area >>>> >>>> >mis-classified > > >>>>is around 41% for T85 CCSM3, 43% for T85 CAM3, and 37% for T42 >>>> >>>> >CAM3. > > >>>>By qualitatively examining the plots you can clearly identify, for >>>>example, two known problems with CCSM climate. The first is that >>>> >>>> >the > > >>>>model is dry over the Amazon (and wet over tropical Africa) and the >>>>Koppen classification indicates that the simulated climate will not >>>>support tropical evergreen trees there. The second is that tundra >>>>extends too far south, due to the cold summer bias in high >>>> >>>> >latitudes > > >>>>(possibly related to low simulated NPP at high latitudes?). >>>> >>>>Anyway, I am proposing this as a potential additional metric to be >>>>provided to the AMWG. It can be used either quantitatively or >>>>qualitatively and is an efficient means of identifying issues with >>>> >>>> >the > > >>>>simulated climate that are directly relevant to carbon cycle >>>> >>>> >modeling. > > >>>>Regards, >>>> >>>>Dave >>>> >>>>Koppen classification rules: >>>> >>>>A Tropical rainy climate in which the temperature of the >>>> >>>> >coolest > > >>>>month is greater than 18C >>>>Af Rainfall in driest month exceeds 6cm. Typical veg: >>>> >>>> >evergreen > > >>>>tropical rainforest >>>>Aw Rainfall in driest month is less than 6cm. Typical veg: >>>>tropical deciduous forests and savannas >>>> >>>>B Dry climates. These climates occur where R, the annual >>>>rainfall in cm is less than the amounts given by the following >>>> >>>> >formulas > > >>>> R = 2T if the rainfall max occurs in winter (the wettest >>>>winter month has more than 3x rainfall as driest summer month) >>>> R = 2T + 7 if rainfall is evenly distributed throughout the >>>> >>>> >year > > >>>> R = 2T + 14 if rainfall max occurs in summer (the wettest >>>> >>>> >summer > > >>>>month has more than 10x the rainfall as the driest winter month) >>>>BS Semiarid climate. Rainfall exceeds one-half the amount >>>> >>>> >given > > >>by >> >> >>>>the formulas for teh boundary between dry and hmid climatees. >>>> >>>> >Typical > > >>>>veg: grasslands, dry tropical savannas, low shrubs >>>>BW Arid climates. Rainfall is less than one half the amount >>>> >>>> >>given >> >> >>>>by the formulas for the boundary between dry and humid climates. >>>>Typical veg: very sparse or lacking >>>> >>>>C Warm temperate rainy climates. The temp of the coolest >>>> >>>> >month > > >>>>is between 18 and -3 C >>>>Cs Summer drought. Rainfall of wettest winter month is at >>>> >>>> >least > > >>>>3x that of driest summer month. Typical veg: trees adapted to >>>> >>>> >summer > > >>>>dryness such as oaks, ecualyptus or olives >>>>Cw Winter drought. Rainfall of wettest summer month is 10x >>>> >>>> >that > > >>>>of driesst winter month. >>>>Cf No drought period. Typical veg: deciduous forests, >>>> >>>> >>subtropical >> >> >>>>evergreen forests. >>>> >>>>D Boreal climate. Temp of the coolest month is less than >>>> >>>> >-3C > > >>and >> >> >>>>temp of warmest month is greater than 10C. Typical veg: Evergreen >>>>coniferous forest, mixed deciduous and evergreen forests near >>>> >>>> >southern > > >>>>boundary >>>>Dw Winter drought. Rainfall of wettest summer month is at >>>> >>>> >least > > >>>>10x that of driest winter month >>>>Df No drought period. >>>> >>>>E Snow climates. Temp of warmest month is less than 10C >>>>ET Tundra. Temp of warmest month is greater than 0C. Typical >>>> >>>> >>veg: >> >> >>>>Tundra, no trees >>>>EF Perpetual frost. Temp of warmest month is less than 0C >>>> >>>>Order of classification E, B, A, C, D. For B, C, D climates summer >>>> >>>> >is > > >>>>warmest 3 months of year, winter is coldest 3 months of year. >>>> >>>> >>>> >>>> >>----------------------------------------------------------------------- >>- >> >> >>>> >>>> >>----------------------------------------------------------------------- >>- >> >> >>----------------------------------------------------------------------- >>- >> >> >>>>_______________________________________________ >>>>CCSM-Land mailing list >>>>CCSM-Land@cgd.ucar.edu >>>>http://www.cgd.ucar.edu/mailman/listinfo/ccsm-land >>>> >>>> >>>> >>>_______________________________________________ >>>CCSM-Land mailing list >>>CCSM-Land@cgd.ucar.edu >>>http://www.cgd.ucar.edu/mailman/listinfo/ccsm-land >>> >>> >>_______________________________________________ >>CCSM-Biogeochemistry mailing list >>CCSM-Biogeochemistry@cgd.ucar.edu >>http://www.cgd.ucar.edu/mailman/listinfo/ccsm-biogeochemistry >> >> > >_______________________________________________ >CCSM-Biogeochemistry mailing list >CCSM-Biogeochemistry@cgd.ucar.edu >http://www.cgd.ucar.edu/mailman/listinfo/ccsm-biogeochemistry > > > -- Eric T. Sundquist U. S. Geological Survey 384 Woods Hole Road Quissett Campus Woods Hole, Massachusetts 02543 USA Phone 508-457-2397 Fax 508-457-2310 -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.cgd.ucar.edu/pipermail/ccsm-biogeochemistry/attachments/20060905/b565689f/attachment.html From kingaw at ornl.gov Tue Sep 5 12:39:58 2006 From: kingaw at ornl.gov (King, Anthony Wayne) Date: Tue Sep 5 13:07:37 2006 Subject: [CCSM-Biogeochemistry] AGU Session: Carbon and Energy Management Analyses Using Coupled Climate-Carbon Models Message-ID: A reminder to all climate-carbon folks that the abstract deadline of September 7 for the Fall AGU is only a couple of days away., We invite you to submit an abstract through the AGU website http://www.agu.org/meetings/fm06/ for the Fall AGU session: B18: Carbon and Energy Management Analyses Using Coupled Climate-Carbon Models We would like to solicit submission of any relevant modeling or observational studies looking at climate-carbon modeling and interactions. Please also note that we encourage any policy-relevant studies. You can find the full session abstract below. The deadline for submission to this year's AGU is September 7. Please forward this to any interested parties. Session description: Future climate and atmospheric conditions are likely to alter the distribution and balance of carbon sources and sinks on both land and in the ocean. Climate change will also likely influence future energy use both in response to demands for heating and cooling and as energy-use and greenhouse gas emission policy and practice respond or adapt to changes in climate. The conventional approach to simulating this influence is to use scenarios of changes in atmospheric CO2 and climate as external inputs to energy and carbon cycle models. However, this approach decouples potentially important feedbacks between the carbon cycle and climate and between climate and energy use and policy, and contributes uncertainty to the simulation of future carbon sequestration and the evaluation of carbon management options. Simulation of carbon management in coupled climate-carbon models avoids this uncertainty. Simulations with general circulation models that include an interactive global carbon cycle indicate a positive feedback between climate change and atmospheric CO2 concentration. Future CO2 concentrations are higher and temperature increases are larger in the coupled climate-carbon cycle simulations than in simulations without the coupling. Investigation of emission reductions, carbon sequestration and carbon management in coupled climate-carbon models thus not only provides the climate resulting from a particular future carbon cycle for input to the carbon management simulations, but also provides for a direct prognostic simulation of future climate change resulting from specific scenarios of carbon management. This session is focused on investigations of carbon management in coupled climate-carbon models, with carbon management broadly defined to include specific carbon sequestration practices such as reduced tillage to sequester carbon in agricultural soils as well as reductions in carbon emissions resulting from energy-use policy and practice. Looking forward to seeing you in San Francisco! Best, Dave Erickson and Tony King ********************************************* David J. Erickson III, PhD Climate and Carbon Research Institute Center for Computational Sciences Oak Ridge National Laboratory 1 Bethel Valley Rd. (overnight mail) PO Box 2008, MS 6016 Oak Ridge, TN 37831-6016 Phone: 865-574-3136 Fax: 865-576-5491 ericksondj@ornl.gov Anthony W. King Environmental Sciences Division Building 1509, MS 6335 Oak Ridge National Laboratory PO Box 2008 Oak Ridge, TN 37831-6335 Phone: 865-576-3436 Fax: 865-574-2232 E-mail: kingaw@ornl.gov -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.cgd.ucar.edu/pipermail/ccsm-biogeochemistry/attachments/20060905/cc22c222/attachment-0001.html From pjr at ucar.edu Tue Sep 5 13:35:43 2006 From: pjr at ucar.edu (Phil Rasch) Date: Tue Sep 5 14:53:14 2006 Subject: [CCSM-Biogeochemistry] Re: [CCSM-LandWG] Physical component biases and BGC: Koppen climate-vegetation classification as a metric In-Reply-To: <44FDB464.2020307@usgs.gov> References: <44FDB464.2020307@usgs.gov> Message-ID: <20060905193543.GA27687@ucar.edu> Hi All, Please see the comments below. Eric said .... > > There is certainly a need for quick vegetation-relevant metrics that can > be applied to the physical climate model. Simple land-area > mis-classificiation metrics treat all mismatches as equally important. > Perhaps a next step would be to evaluate the extent to which mismatches > would be critical for climate/veg/carbon feedbacks -- for example, those > that might have a lot of leverage on biomass, ET, and/or albedo. It > wouldn't be difficult to quantify these. This approach would probably > require adding PET or an aridity index (a la Holdridge -- I can hear Dave > S groaning now!). > > Eric > It is gratifying to see the amount of interest that is indicated by this email traffic (some of it offline as well). As Dave Lawrence said (in an offline email to a subset of us) > > I think that perhaps the point is being missed just a bit. The issue is > that the AMWG is looking for guidance as to what biases in the simulated > climate are the most harmful to the vegetation simulation when run in > DGVM/Carbon cycle mode. The point of a Koppen-like classification, even > if it is old (although recently modified), is that it can direct efforts > to reduce biases and serves as an efficient means to check whether or > not the efforts are having any success without running the DGVM or > carbon cycle models. A Koppen classification is not meant as a strict > check on the simulated climate. It is a model development tool, and > undoubtedly an imperfect one, but possibly better or at least different > to the standard method which is to simply compare maps and climate > statistics against observed temperature and precipitation climatologies. To be even more AMWG-centric. We need a set of "numbers" that tell us whether the model that we are producing is going to "improve" simulations produced with a DGVM or Carbon Cycle model or not. Although we will of course ask the BGC and Land Working groups to look at promising simulations, it will help AMWG enormously if we could produce a few (less than 10) numbers for any simulation, that will tell the developers whether we are on the right track or not, even for simulations that we would not want to ask the broader community to scrutinize. Ideally the numbers would provide us with a quantitative statement of whether any given simulation is is better, the same as, or worse than a control. By "the same" I mean we want some indication of how much different two cases have to be to in order to be perceived as "different". You guys are welcome to produce whichever metrics you think are best (and achievable). For the moment, it is clear that Dave's Koppen codes are good start. Phil -- Phil Rasch, Climate Modeling Section, National Center for Atmospheric Research Mail --> P.O. Box 3000, Boulder CO 80307 Shipping --> 1850 Table Mesa Dr, Boulder, CO 80305 email: pjr@ucar.edu, Web: http://www.cgd.ucar.edu/cms/pjr Phone: 303-497-1368, FAX: 303-497-1324 From robted at eas.gatech.edu Tue Sep 5 14:47:00 2006 From: robted at eas.gatech.edu (Robert Dickinson) Date: Tue Sep 5 14:53:14 2006 Subject: [CCSM-Biogeochemistry] Re: [CCSM-LandWG] Physical component biases and BGC: Koppen climate-vegetation classification as a metric In-Reply-To: <20060903175106.GA3272@ucar.edu> References: <44F730FB.1000008@ucar.edu> <44F8B6D7.7090400@ucar.edu> <20060903175106.GA3272@ucar.edu> Message-ID: <6.1.0.6.2.20060905164450.02a5bf78@pop.mail.gatech.edu> An HTML attachment was scrubbed... URL: http://www.cgd.ucar.edu/pipermail/ccsm-biogeochemistry/attachments/20060905/51f727d0/attachment.html From forrest at climate.ornl.gov Tue Sep 5 21:36:26 2006 From: forrest at climate.ornl.gov (Forrest Hoffman) Date: Wed Sep 6 07:03:54 2006 Subject: [CCSM-Biogeochemistry] Re: [CCSM-LandWG] Physical component biases and BGC: Koppen climate-vegetation classification as a metric In-Reply-To: <6.1.0.6.2.20060905164450.02a5bf78@pop.mail.gatech.edu> References: <44F730FB.1000008@ucar.edu> <44F8B6D7.7090400@ucar.edu> <20060903175106.GA3272@ucar.edu> <6.1.0.6.2.20060905164450.02a5bf78@pop.mail.gatech.edu> Message-ID: <44FE423A.1080205@climate.ornl.gov> K?ppen is an expert-derived classification of low dimensionality (bi-variate) that has been around for a long time. Newer alternatives, like Multivariate Spatio-Temporal Clustering (MSTC), provide objective, repeatable, and fully quantitative methods of regionalization or discretization with arbitrary dimensionality. Moreover, the regimes produced have nearly equal multi-variance (with respect to normalized variables). Since the classification is quantitative, it may be used to intercompare various model results and/or observational datasets irrespective of native resolution; and "similarity" and "representativeness" between regimes can be computed across datasets. For example, see Hoffman, Forrest M., William W. Hargrove, David J. Erickson, and Robert J. Oglesby. August 3, 2005. "Using Clustered Climate Regimes to Analyze and Compare Predictions from Fully Coupled General Circulation Models ." /Earth Interactions/ , *9*(10): 1-27. doi:10.1175/EI110.1 . [ PDF ] In the present context, MSTC could be used to calculate pre-determined regimes based on a suite of variables from Willmott and Matsuura, then the resulting centroids would be applied in a single-pass classification to model output. One can then compute a degree of similarity/difference between maps. It is straightforward to then figure out which axes (or model output variables) are the largest contributors to the difference between classification results. Forrest Robert Dickinson wrote: > Ann Henderson Sellers was working with the Koppen classification and > some version of CCM about 2 decades ago. She made it into the first > dynamic vegetation GCM if I am remembering correctly > > Bob Dickinson > > At 01:51 PM 9/3/2006, Phil Rasch wrote: > >> Its a step forward if we have been building models without comparing >> to something like this before! >> >> Phil >> >> On Fri, Sep 01, 2006 at 04:40:23PM -0600, David Schimel wrote: >> > we are going to evaluate a third generation climate model against >> 1930s >> > biogeography? What a step forward! >> > >> > David Lawrence wrote: >> > >Hi everyone, >> > > >> > >At the Aspen Global Change Institute meeting on ESMs a few weeks ago >> > >there was discussion of metrics for this new class of model. A metric >> > >that GFDL is using (or intends to use) is the Koppen >> > >climate-vegetation classification. The Koppen climate classification >> > >characterizes 11 distinct seasonal temperature and precipitation >> > >regimes and their associated vegetation classes (see classification >> > >rules below). It is an efficient means of identifying climate biases >> > >that are likely to affect carbon cycle modeling. >> > > >> > >I coded up the classification procedure (note that I am using the >> > >modified Koppen classification by Guetter and Kutzback, 1990) and ran >> > >it through some CCSM data as a first check to see whether or not it >> > >could be useful in terms of directing model development efforts. >> > > >> > >I attach two plots that were created using CCSM3 data (T85 fully >> > >coupled and AMIP, and T42 AMIP). The percent land area >> mis-classified >> > >is around 41% for T85 CCSM3, 43% for T85 CAM3, and 37% for T42 CAM3. >> > >By qualitatively examining the plots you can clearly identify, for >> > >example, two known problems with CCSM climate. The first is that the >> > >model is dry over the Amazon (and wet over tropical Africa) and the >> > >Koppen classification indicates that the simulated climate will not >> > >support tropical evergreen trees there. The second is that tundra >> > >extends too far south, due to the cold summer bias in high latitudes >> > >(possibly related to low simulated NPP at high latitudes?). >> > > >> > >Anyway, I am proposing this as a potential additional metric to be >> > >provided to the AMWG. It can be used either quantitatively or >> > >qualitatively and is an efficient means of identifying issues with >> the >> > >simulated climate that are directly relevant to carbon cycle modeling. >> > > >> > >Regards, >> > > >> > >Dave >> > > >> > >Koppen classification rules: >> > > >> > >A Tropical rainy climate in which the temperature of the >> coolest >> > >month is greater than 18C >> > >Af Rainfall in driest month exceeds 6cm. Typical veg: evergreen >> > >tropical rainforest >> > >Aw Rainfall in driest month is less than 6cm. Typical veg: >> > >tropical deciduous forests and savannas >> > > >> > >B Dry climates. These climates occur where R, the annual >> > >rainfall in cm is less than the amounts given by the following >> formulas >> > > R = 2T if the rainfall max occurs in winter (the wettest >> > >winter month has more than 3x rainfall as driest summer month) >> > > R = 2T + 7 if rainfall is evenly distributed throughout the >> year >> > > R = 2T + 14 if rainfall max occurs in summer (the wettest summer >> > >month has more than 10x the rainfall as the driest winter month) >> > >BS Semiarid climate. Rainfall exceeds one-half the amount >> given by >> > >the formulas for teh boundary between dry and hmid climatees. Typical >> > >veg: grasslands, dry tropical savannas, low shrubs >> > >BW Arid climates. Rainfall is less than one half the amount >> given >> > >by the formulas for the boundary between dry and humid climates. >> > >Typical veg: very sparse or lacking >> > > >> > >C Warm temperate rainy climates. The temp of the coolest month >> > >is between 18 and -3 C >> > >Cs Summer drought. Rainfall of wettest winter month is at least >> > >3x that of driest summer month. Typical veg: trees adapted to summer >> > >dryness such as oaks, ecualyptus or olives >> > >Cw Winter drought. Rainfall of wettest summer month is 10x that >> > >of driesst winter month. >> > >Cf No drought period. Typical veg: deciduous forests, >> subtropical >> > >evergreen forests. >> > > >> > >D Boreal climate. Temp of the coolest month is less than >> -3C and >> > >temp of warmest month is greater than 10C. Typical veg: Evergreen >> > >coniferous forest, mixed deciduous and evergreen forests near southern >> > >boundary >> > >Dw Winter drought. Rainfall of wettest summer month is at least >> > >10x that of driest winter month >> > >Df No drought period. >> > > >> > >E Snow climates. Temp of warmest month is less than 10C >> > >ET Tundra. Temp of warmest month is greater than 0C. >> Typical veg: >> > >Tundra, no trees >> > >EF Perpetual frost. Temp of warmest month is less than 0C >> > > >> > >Order of classification E, B, A, C, D. For B, C, D climates summer is >> > >warmest 3 months of year, winter is coldest 3 months of year. >> > > >> > > >> > >> >------------------------------------------------------------------------ >> > > >> > > >> > >> >------------------------------------------------------------------------ >> > > >> > >> >------------------------------------------------------------------------ >> > > >> > >_______________________________________________ >> > >CCSM-Land mailing list >> > >CCSM-Land@cgd.ucar.edu >> > >http://www.cgd.ucar.edu/mailman/listinfo/ccsm-land >> > > >> > _______________________________________________ >> > CCSM-Land mailing list >> > CCSM-Land@cgd.ucar.edu >> > http://www.cgd.ucar.edu/mailman/listinfo/ccsm-land >> _______________________________________________ >> CCSM-Biogeochemistry mailing list >> CCSM-Biogeochemistry@cgd.ucar.edu >> http://www.cgd.ucar.edu/mailman/listinfo/ccsm-biogeochemistry > >------------------------------------------------------------------------ > >_______________________________________________ >CCSM-Land mailing list >CCSM-Land@cgd.ucar.edu >http://www.cgd.ucar.edu/mailman/listinfo/ccsm-land > > -- Forrest Hoffman mailto:forrest@climate.ornl.gov Oak Ridge National Laboratory mailto:forrest@computer.org Computer Science & Mathematics http://climate.ornl.gov/~forrest and Environmental Sciences Divisions (865) 576-7680 voice Building 5600, Room C221, MS 6016 (865) 576-5491 fax P.O. Box 2008 Deliveries: Bethel Valley Road Oak Ridge TN 37831-6016 35? 55' 23" N 84? 19' 20" W "The ability to speak doesn't Check out ORNL's make you intelligent." Climate & Carbon Research Institute -- Qui-Gon Jinn at http://www.ccs.ornl.gov/CCR/ -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.cgd.ucar.edu/pipermail/ccsm-biogeochemistry/attachments/20060905/a7350c6c/attachment-0001.html From ericksondj at ornl.gov Wed Sep 6 08:44:39 2006 From: ericksondj at ornl.gov (Erickson III, David J.) Date: Wed Sep 6 10:04:31 2006 Subject: [CCSM-Biogeochemistry] Re: [CCSM-LandWG] Physical component biasesand BGC: Koppen climate-vegetation classification as a metric References: <44F730FB.1000008@ucar.edu> <44F8B6D7.7090400@ucar.edu> <20060903175106.GA3272@ucar.edu> <6.1.0.6.2.20060905164450.02a5bf78@pop.mail.gatech.edu> <44FE423A.1080205@climate.ornl.gov> Message-ID: <537C6C0940C6C143AA46A88946B85417042C2EC4@ORNLEXCHANGE.ornl.gov> The MSTC approach Forrest describes is far more quantitative and usefull than trying to eyeball climate statistics from Koppen. Perhaps we could be very specific about the anticipated (and unantcipated!) biaises in the CCSM or whatever model with centroids computed from Willmont et al.. The Hoffman et al paper outlines this approach.. dave ________________________________ From: ccsm-biogeochemistry-bounces@cgd.ucar.edu on behalf of Forrest Hoffman Sent: Tue 9/5/2006 11:36 PM To: Robert Dickinson Cc: ccsm-biogeochemistry@cgd.ucar.edu; ccsm-land@cgd.ucar.edu; David Schimel; pjr@ucar.edu Subject: Re: [CCSM-Biogeochemistry] Re: [CCSM-LandWG] Physical component biasesand BGC: Koppen climate-vegetation classification as a metric K?ppen is an expert-derived classification of low dimensionality (bi-variate) that has been around for a long time. Newer alternatives, like Multivariate Spatio-Temporal Clustering (MSTC), provide objective, repeatable, and fully quantitative methods of regionalization or discretization with arbitrary dimensionality. Moreover, the regimes produced have nearly equal multi-variance (with respect to normalized variables). Since the classification is quantitative, it may be used to intercompare various model results and/or observational datasets irrespective of native resolution; and "similarity" and "representativeness" between regimes can be computed across datasets. For example, see Hoffman, Forrest M., William W. Hargrove, David J. Erickson, and Robert J. Oglesby. August 3, 2005. "Using Clustered Climate Regimes to Analyze and Compare Predictions from Fully Coupled General Circulation Models ." Earth Interactions , 9(10): 1-27. doi:10.1175/EI110.1 . [ PDF ] In the present context, MSTC could be used to calculate pre-determined regimes based on a suite of variables from Willmott and Matsuura, then the resulting centroids would be applied in a single-pass classification to model output. One can then compute a degree of similarity/difference between maps. It is straightforward to then figure out which axes (or model output variables) are the largest contributors to the difference between classification results. Forrest Robert Dickinson wrote: Ann Henderson Sellers was working with the Koppen classification and some version of CCM about 2 decades ago. She made it into the first dynamic vegetation GCM if I am remembering correctly Bob Dickinson At 01:51 PM 9/3/2006, Phil Rasch wrote: Its a step forward if we have been building models without comparing to something like this before! Phil On Fri, Sep 01, 2006 at 04:40:23PM -0600, David Schimel wrote: > we are going to evaluate a third generation climate model against 1930s > biogeography? What a step forward! > > David Lawrence wrote: > >Hi everyone, > > > >At the Aspen Global Change Institute meeting on ESMs a few weeks ago > >there was discussion of metrics for this new class of model. A metric > >that GFDL is using (or intends to use) is the Koppen > >climate-vegetation classification. The Koppen climate classification > >characterizes 11 distinct seasonal temperature and precipitation > >regimes and their associated vegetation classes (see classification > >rules below). It is an efficient means of identifying climate biases > >that are likely to affect carbon cycle modeling. > > > >I coded up the classification procedure (note that I am using the > >modified Koppen classification by Guetter and Kutzback, 1990) and ran > >it through some CCSM data as a first check to see whether or not it > >could be useful in terms of directing model development efforts. > > > >I attach two plots that were created using CCSM3 data (T85 fully > >coupled and AMIP, and T42 AMIP). The percent land area mis-classified > >is around 41% for T85 CCSM3, 43% for T85 CAM3, and 37% for T42 CAM3. > >By qualitatively examining the plots you can clearly identify, for > >example, two known problems with CCSM climate. The first is that the > >model is dry over the Amazon (and wet over tropical Africa) and the > >Koppen classification indicates that the simulated climate will not > >support tropical evergreen trees there. The second is that tundra > >extends too far south, due to the cold summer bias in high latitudes > >(possibly related to low simulated NPP at high latitudes?). > > > >Anyway, I am proposing this as a potential additional metric to be > >provided to the AMWG. It can be used either quantitatively or > >qualitatively and is an efficient means of identifying issues with the > >simulated climate that are directly relevant to carbon cycle modeling. > > > >Regards, > > > >Dave > > > >Koppen classification rules: > > > >A Tropical rainy climate in which the temperature of the coolest > >month is greater than 18C > >Af Rainfall in driest month exceeds 6cm. Typical veg: evergreen > >tropical rainforest > >Aw Rainfall in driest month is less than 6cm. Typical veg: > >tropical deciduous forests and savannas > > > >B Dry climates. These climates occur where R, the annual > >rainfall in cm is less than the amounts given by the following formulas > > R = 2T if the rainfall max occurs in winter (the wettest > >winter month has more than 3x rainfall as driest summer month) > > R = 2T + 7 if rainfall is evenly distributed throughout the year > > R = 2T + 14 if rainfall max occurs in summer (the wettest summer > >month has more than 10x the rainfall as the driest winter month) > >BS Semiarid climate. Rainfall exceeds one-half the amount given by > >the formulas for teh boundary between dry and hmid climatees. Typical > >veg: grasslands, dry tropical savannas, low shrubs > >BW Arid climates. Rainfall is less than one half the amount given > >by the formulas for the boundary between dry and humid climates. > >Typical veg: very sparse or lacking > > > >C Warm temperate rainy climates. The temp of the coolest month > >is between 18 and -3 C > >Cs Summer drought. Rainfall of wettest winter month is at least > >3x that of driest summer month. Typical veg: trees adapted to summer > >dryness such as oaks, ecualyptus or olives > >Cw Winter drought. Rainfall of wettest summer month is 10x that > >of driesst winter month. > >Cf No drought period. Typical veg: deciduous forests, subtropical > >evergreen forests. > > > >D Boreal climate. Temp of the coolest month is less than -3C and > >temp of warmest month is greater than 10C. Typical veg: Evergreen > >coniferous forest, mixed deciduous and evergreen forests near southern > >boundary > >Dw Winter drought. Rainfall of wettest summer month is at least > >10x that of driest winter month > >Df No drought period. > > > >E Snow climates. Temp of warmest month is less than 10C > >ET Tundra. Temp of warmest month is greater than 0C. Typical veg: > >Tundra, no trees > >EF Perpetual frost. Temp of warmest month is less than 0C > > > >Order of classification E, B, A, C, D. For B, C, D climates summer is > >warmest 3 months of year, winter is coldest 3 months of year. > > > > > >------------------------------------------------------------------------ > > > > > >------------------------------------------------------------------------ > > > >------------------------------------------------------------------------ > > > >_______________________________________________ > >CCSM-Land mailing list > >CCSM-Land@cgd.ucar.edu > >http://www.cgd.ucar.edu/mailman/listinfo/ccsm-land > > > _______________________________________________ > CCSM-Land mailing list > CCSM-Land@cgd.ucar.edu > http://www.cgd.ucar.edu/mailman/listinfo/ccsm-land _______________________________________________ CCSM-Biogeochemistry mailing list CCSM-Biogeochemistry@cgd.ucar.edu http://www.cgd.ucar.edu/mailman/listinfo/ccsm-biogeochemistry ________________________________ _______________________________________________ CCSM-Land mailing list CCSM-Land@cgd.ucar.edu http://www.cgd.ucar.edu/mailman/listinfo/ccsm-land -- Forrest Hoffman mailto:forrest@climate.ornl.gov Oak Ridge National Laboratory mailto:forrest@computer.org Computer Science & Mathematics http://climate.ornl.gov/~forrest and Environmental Sciences Divisions (865) 576-7680 voice Building 5600, Room C221, MS 6016 (865) 576-5491 fax P.O. Box 2008 Deliveries: Bethel Valley Road Oak Ridge TN 37831-6016 35? 55' 23" N 84? 19' 20" W "The ability to speak doesn't Check out ORNL's make you intelligent." Climate & Carbon Research Institute -- Qui-Gon Jinn at http://www.ccs.ornl.gov/CCR/