Second GREENCYCLES MC-RTN Network Meeting Minutes

15th - 18th February, 2006
Lisbon, Portugal

Host: José M.C. Pereira, DEF, Instituto Superior de Agronomia, Tapada da Ajuda 1349-017, Lisboa

Meeting Venue: Vila Galé – Hotel Opera

Part 2:
       
  Science objective 2: Determine the effect of land use on climate  
       
  Title: Effect of land use policies and agricultural management strategies on the global biogeochemical cycles: Simulation with the LPJ dynamic global vegetation model,
ESRIII: Marlies
 
    Supervisors: Alberte Bondeau, Wolfgang Cramer  
       
    Aims: better representation of agricultural management in model and quantifying the role of different practices and policies on biospheric processes. Only simple land management functions in mLPJ, want better to quantify the effects of land use. Characterise worldwide. Use of fire in agriculture. Nice to have N cycle. Implement managed biosphere.  
       
  Sergey: data to quantify agricultural fires?  
  Marlies: publications, very general statistics (e.g. county level in the US, but very heterogeneous). Collaboration with ISA.  
  Jose: different data sets (landuse and fire) from satellite  
  Andrew: You compile a global data base on management – incredibly ambitious.  
  Alberte: true but inevitable.  
  Andrew: should be constrained to a certain amount of work, otherwise too much.  
  Sandy: how does this relate LCCS (FAO)? LCCS provides a hierarchical classification of what is an important management system.  
  Alberte: possible to have a demand for more data than that.  
  Sandy: under LUCC there is an activity to feed into LCCS, there is a potential for feedback from Marlies research into the LUCC project topic 1.  
  Jose: JRC global landcover initiative conforms to LCCS, but more regional detail.  
 
 
Science Objective 3: Improve understanding of natural sources of CH4 and their responses to human activities
 
Title: Methane processes and controls,
ESR X: Marcin
    Supervisor: Torben Christensen, ULUND  
       
  Sergey: How will you be scaling your result up to smaller scales?
 
  Sandy : Secondments at MetO, and UBRIS will help to achieve this.  
  Han: Link becomes clear after Roxana’s talk.  
       
  Title: Hydrology and scaling,
ESR XI: Roxana
    Supervisor: Han Dolman, VAU  
       
  Torben: There is a lot of outside interest in this area, also outside GC. Should have outside workshop on this.  
  Allan: Will tropical peatlands be dealt with as well?  
  Torben: lacking data.  
  Allan: but large emissions very high from there.  
  Sandy: nonetheless UBRIS is developing global methane model…  
       
    [see also discussion on methane workshop below]  
       
  Title: Changing CH4 and CO global budgets: New constraints from their mixing and stable isotopic ratios measured in ice cores,
ESR XII Anna
    Supervisor: Jérôme Chappellaz, LGGE  
       
  Sandy: secondments at MetOffice and UBRIS should be the coordinated and organised as one single activity.  
       
 
   
  Science Objective 4: Quantifying impacts of climate change and climate variability on fire-induced emissions of greenhouse gases
       
  Title: Land Use and Fire: Global Greenhouse Gas Budgets,
ER II, Allen
       
  Nicolas: What do estimate as the total global emissions from biomass
burning from wildland fires?		  
  Allen: We estimate that between 2-4 Pg per year are emitted from above
ground fires using LPJ+SPITFIRE. Extrapolating from emission estimates from
peat fires in Borneo to form a global estimate from below-ground fires, one
could add a further 1-2 Pg to that figure. It is very difficult to assess
the accuracy of these figures given that there are no other comparative
studies available. Furthermore, the satellite monitoring record for
greenhouse trace gas emissions and related work on inverse modelling of
carbon sources are still in their relative infancy.		  
  Jose: What is the function you use to describe human-caused
ignitions in LPJ+SPITFIRE?		  
  Allen: Shows graph. It is a non-linear function that attempts to capture a)
fire suppression in highly urbanised areas , and ii) land use or lifestyle
characteristics of regional inhabitants- the aNd parameter which we
calibrate using regional (not global) observed fire datsets, where possible.		  
       
       
  Title: Global fire activity and its connection with the climate system,
ESRIX Yannick
    Supervisor: Jose Pereira, ISA  
       
  Sandy/Kirsten: it may be a problem to relate fire activities observed by satellites to climatic modes such as ENSO because of the long timescale of ENSO and the comparable short timescale of satellite data – probably only covering one or two full cycles.  
  Andrew Could use climate reanalysis (such as ERA40) to obtain fields of soil moisture and precipitation to evaluate remotely sensed fire patterns.  
  ?? RS may be wrong tool for looking at modes of variability. May be hard to relate to ENSO.  
  Allen: If the analysis is restricted to the tropics and
sub-tropics only (shorter fire return intervals), then this should yield
sufficient degrees of freedom to examine ENSO effects (in terms of intensity
and duration) on fire activity. This and the fact that his fire data are at
monthly resolution, not yearly, should obviate the concern of having an
insufficiently long time series to undertake his project.		  
       
  Thursday, February 16th  
     
 
       
  Science Objective 5: Quantify impacts of climate change on terrestrial and oceanic biogenic emissions of aerosols and chemically active gases and their effects on tropospheric chemistry
       
  ESRII (reported by Stephen)
   

Supervisor Mike Sanderson, Richard Betts

Hadley uses Guenther approach (1993) at the moment to calculate BVOC emissions, CO2 interactions are not taken account of. ESRII should update these effects applying new dependencies to CO2 etc.

3 out of 5 candidates have already been interviewed. Start date April, but subject to the time required for security clearance.

MetOffice are confident to be funded for the period after GC to compensate for the last start. Secondments planned in Bialstok, Poland (Stephen has not been able to get in touch with them yet), and at the University Lund (Almut Arneth).
Possible collaboration with Trevor Keenan (ESRVII)

 
       
  Title: Marine Production of DMS and its Interaction with Climate,
ESR V: Meike
    Supervisor: Corrine Le Queré, Peter Liss, UEA  
       
  Andrew: Is somebody working on the feedback of DMS on climate (not the otherway round)?  
  Meike: Coupling seems not possible for thesis, doing offline forcing. There are some model results that show locally ~2deg C change, but globally the feedback seems to be not so important.  
   
 
       
  Science Objective 6: Quantify impacts of vegetation and climate changes on atmospheric dust and its feedbacks to CO2 and climate
       
  Title: Global dust processes and consequences,
ESR VII: Sarah
    Supervisor: Sandy Harrison and Colin Prentice, UBRIS  
       
  Alberte: soil water content important for dust emissions, which soildepth is used?  
  Sandy: should be skin water content – top layer. Explicitly simulate first centimeter? Better soil data would probably not help.  
  Andrew: maybe remote sensed data can help in this case, as they can measure top soil.  
  Dorothea: How could this be evaluated?  
  Andrew: how much work would be necessary to modify LPJ?  
  Sandy: Not too much, possible to do with keeping the focus.  
  Han: Classical example for the importance of sub grid-scale processes. Cannot give advice, but probably rather drastic simplification necessary.  
  Sandy: Got so far as possible without these modification. To get better, gustiness and vegetation spacing are necessary. There is a lot of uncertainty, but no stronger constraint is possible of dust emissions without these processes. The representation may not be completely physical, but perhaps probabilistic.  
       
  Title: Climate forcing by dust and CH4,
ER5: Sandy reports on this position
    Interactions sarah Shannon, ynanick, allen,Pierre (fire) (QUEST rita, wetlands and CH4 + Marcin), validation data from Anna Lourantou…  
       
  Dorothea: How sophisticated are dust models, sources from different soils?  
  Sandy: Predicts where source are from vegetation cover. i) little vegetation equals bare ground, ii) soil types from discover data set, emission from certain topography and fine soil texture (subgrid scale). Take a 5min hydrological model to get basins to find fine grain places within a grid cell. iii) 6hourly wind speed (directly from GCM or corrected by observation)
iv) phenology. This delivers good prediction of present-day emission for Asia and Africa. Also when forced to simulate local individual dust storms the results are good.
  
  Carole: starting date?  
  Sandy: October, which fits good to other schedules. Remote sensing comes in before that, so Sarah and Kirsten might be wanting to use the data for validation.  
  ?? Coupled simulation of climate?  
  Sandy: offline radiative forcing first. To do really coupling need a stable coupled model in either Paris or Exeter.  
       
  Andrew thanks all for their excellent presentation and the stimulating discussions
       
   
  Open discussion of model coupling, led by Pierre and Stephen  
    Presentation of current results from C4MIP and IMOGEN

Conclusion: Models agree on feedback sign, but large uncertainty, less robust in regional response than CO2, need for validation. with forced simulations, models agree (global surface NEE??) within 20-30%, but when coupled with the atmosphere much larger variation. long list of variables to be included into the fully coupled chemistry-climate models

  
  Andrew: Relation to Greencycles: How to better constrain hotspots for ER 3?  
  Pierre: Amazon big unknown (in several models: triffid, lpj, hyland) because of the fate of large C stocks. Also permafrost, frozen carbon methane…Terrestrial Feedbacks are less constrained than ocean. Ocean CO2 uptake is critically controlled by CO2 diffusion into deep ocean and not so much biology, whereas in terrestrial biosphere is different. Higher CO2 may affect ocean biology, but not necessarily ocean carbon fluxes.  
  Andrew: Better constraints due to the work (data/experiments/models) carried out in GC? Work of Raquel, Marcin, Sönke.  
  Han: From CarboEurope there are no manipulation experiments, but for instance the real world experiment of the 2003 summer drought in Europe…?  
  Sönke: Intercomparison on local and regional level done using different models, however, until now only preliminary results.  
       
 
  Open discussion on methane, led by Torben
    Aim: key uncertainties in models and how to meet them best with what is really that Marcin, Roxana and Anna can do to help Sandy, Stephen and Pierre?  
       
  Sergey: several models on methane, what do they need to be improved?  
  Sandy: focus in UBRIS on wetland PFTs, problem: peat body vs. “wetter things” (mangroves and other plants with aerenchyma), leading to preferential emission of CH4) – need several mires of different species composition,  
  Torben: bryophyte ratio as a function of nutrient content. Could models predict that?  
  Sergey: link between CH4 and permafrost?  
  Torben: Our measurements yes, and for example those in Finland, but altogether only few sites.  
  Han: MPI Jena works in the Lena basis (Sergey corrects this to Potsdam)  
  Torben: What about Barrow?  
  Andrew: What’s the greater problem: biogeochemical or hydrology?  
  Han/Torben Hydrology! But it’s difficult, because it’s very small scale.  
  Sandy: issue: scaling, is it possible beyond the 10km scale?  
  Han:
100km upscaling to higher is not so much of a problem, but getting from site level to 10 km is probably a substantial problem.
  
       
    [discussion on secondments at ULUND that took place here moved to the general section on secondments etc. below]. Torben then gives a report on a recent paper by Keppler et al., Nature, suggesting substantial methane emissions on terrestrial plants under aerob conditions.  
       
  Andrew: interest in CH4 stimulated from this paper, good even if science is wrong.
Problem with the review process?
  
  Pierre: when BVOCs are measured, why is CH4 not seen?  
  Torben: those machines don’t see CH4, but probably CH4 is relevant in the VOC affair.  
       
 
  Open discussion on fire, led by Kirsten
    Overview over the QUEST/IGBP fast track initiative fire workshop in Exeter, UK 27-28 Oct. 2005. GC participants: Kirsten, Allen, Sergey, Sandy, Pierre.

Summary: Understanding and representation of fire in carbon cycle currently very active. Observations are emerging to support modelling

Results:
• Synthesis paper on the role of fire in the Earth System (written by Kathy Hibbert),
• Assemblage of paleao data sets to allow hind casting of past fire regime,
• establishment of an information catalogue of fire earth observations products with guidance on the domain of their reliability.

There is a (QUEST-organised) workshop on the palaeo activity in Totnes, UK, 30-Oct. 2-November 2006

Overview on GC fire workshop in Bath (May 2005), protocol available on Greencycles webpage.

Discussion points: Coupling strategy, fire model intercomparison, natural wild fires vs. land-use related fires to obtain global biomass burning C budget estimates. Input to other GC activities: fire quantification for synthesis of other biogeochemistry, CH4, land-use change interactions with nitrogen cycle, dust, ecophysiology in the Mediterranean.

 
       
   

Situation at the Metoffice, Sergey and Stephen

Sergey’s going to work on Triffid to implement an ignition and area burnt model
In parallel ER2 will be implementing an emission sub-model for chemistry and aerosols, using prescribed input of area burnt etc.. Remote sensed products can be used to for model evaluation (example intercomparison with products by JRC and UBRIS, AHVRR based), or used to prescribe area burnt to simulate impacts on atmosphere chemistry and aerosol concentrations.

Sergey’s workplan for the MetOffice:
Key issues: improving models, coupling to GCMs, usage of data for model validation.
• Implementation of fire risk into GCM
• Simplification necessary to be implemented into Hadley model
• Re-parameterisation for PFTs
• Design of fuel functions for PFTs
• Comparison of Moses and Triffid
• Evaluation of existing fire models for their applicability in GCMs.

Data for evaluation: national statistics are not consistent, cosare and potentially biased. RS data gives spatial patterns, but the uncertainty can reach 300% in extreme cases. 

  
       
  ...back