EPOCA WP9: From process studies to ecosystem models Participants involved: LOV, UiB, IFM-GEOMAR, GKSS, KNAW, UGOT, UNIVBRIS (a.o. J.-P. Gattuso, R. Bellerby, M. Schartau, J. Middelburg, A. Oschlies)
Motivation: Current parameterisations of calcification PIC prod. ~ Prim.Prod. (of some PFT, possibly modulated by ) PIC prod. ~ Detritus prod. Essentially all current parameterisations employ Eppley’s temperature dependence.
Calcification & temperature (according to current models) low T low PP, slow microbial loop low PIC prod. high T high PP, fast microbial loop large PIC prod. irrespective of nutrient supply, export production, grazing… low PIC export large PIC export
Example: calcification & temperature UVic model: temperature dependence helps to get latitudinal distribution of rain ratio “right”: (Schmittner et al., 2008)
Example: calcification & temperature Does this give meaningful results in global-warming runs? PICprod PP EP Increase in PIC production closely linked to temperature-driven increase in Prim.Prod. (Schmittner et al., 2008)
General problem with empirical models May work well under empirical conditions No guarantee that this will continue under new environmental conditions –higher temperatures –higher CO 2 –… Aim for mechanistic models
Objectives Integration & Synthesis experimentsmodels Efficient knowledge transfer Feedback to efficiently reduce uncertainty
Approach 1.Analysis experimentsmodels Coherent data base (organisms, ecosystems) Meta-analysis (mesocosm, microcosm) Meta-analysis (model assumptions, parameterisations) T9.1 T9.2 T9.3
Approach 2.Modelling of micro- and mesocosm experiments 2.Model improvement: balance complexity, performance, portability 3.Assessment and recommendations for incorporation into global-scale models experimentsmodels Data-assimilative parameter estimation T9.4 T9.5 T9.6
Deliverables D9.1: advice/guidance: data storage/documentation/protocol (month 2, R, PU) D9.2: structured data base (month 12, R, PP) D9.3: Mesocosm meta-analysis, guidance to future experiments (month 12, R, PP) D9.4: Identification of physiological/ecological processes that contribute most to uncertainties in ecosystem models (month 24, R, PU) D9.5: Improved model formulation for pH-sensitive processes -> Earth system models (month 40, R, PU) D9.6: Uncertainty analysis (month 48, R, PU)
Example 1 Calibration by chemostat/turbidostat data (Pahlow & Oschlies, subm.) Chain model of N, P, light colimitation
Example 2 Calibration by mesocosm data (Schartau et al., 2007)
Example 3: Transfer to global models 350 ppm 700 ppm 1050 ppm (Riebesell et al., 2007)(Oschlies et al., subm.) 50% increase in suboxic volume (<5mmol/m 3 )
Questions from model study & feedback to experimentalists Temperature effects vs. pH effects? Observational evidence of pCO 2 -sensitive C:N ratios in the ocean? What is the mechanism for export of excess C?