12
Der Ozean im globalen Wandel: wärmer, saurer, atemlos Ulf Riebesell GEOMAR HelmholtzZentrum für Ozeanforschung Kiel FONAForum, Leipzig, 9.11. 09. 2013

Der$Ozean$im$globalen$Wandel:$ wärmer,$saurer,$atemlos$ · wärmer$ saurer$ atemlos$ sinkende$Produk]vität steigende$Produk]vität Gruber2011 korrosivfür$ Aragonit Sauerstoffmangel$

Embed Size (px)

Citation preview

Page 1: Der$Ozean$im$globalen$Wandel:$ wärmer,$saurer,$atemlos$ · wärmer$ saurer$ atemlos$ sinkende$Produk]vität steigende$Produk]vität Gruber2011 korrosivfür$ Aragonit Sauerstoffmangel$

Der  Ozean  im  globalen  Wandel:  wärmer,  saurer,  atemlos  

 Ulf  Riebesell  

GEOMAR  Helmholtz-­‐Zentrum  für  Ozeanforschung  Kiel    

FONA-­‐Forum,  Leipzig,  9.-­‐11.  09.  2013  

Page 2: Der$Ozean$im$globalen$Wandel:$ wärmer,$saurer,$atemlos$ · wärmer$ saurer$ atemlos$ sinkende$Produk]vität steigende$Produk]vität Gruber2011 korrosivfür$ Aragonit Sauerstoffmangel$

wärmer  Oberflächentemperatur  

saurer  pH  Wert  

atemlos  Sauerstoffgehalt  

Der  Ozean    im  globalen  Wandel:  

Historisch  RCP2.6  

Stabilisierung    bei  450  ppm  CO2  

RCP8.5  „business-­‐as-­‐usual“  

FONA-­‐Forum,  Leipzig,  9.-­‐11.  09.  2013  

Bopp  et  al.  2013  

0  

Page 3: Der$Ozean$im$globalen$Wandel:$ wärmer,$saurer,$atemlos$ · wärmer$ saurer$ atemlos$ sinkende$Produk]vität steigende$Produk]vität Gruber2011 korrosivfür$ Aragonit Sauerstoffmangel$

wärmer   saurer   atemlos  sinkende  Produk]vität  

steigende  Produk]vität  Gruber  2011  

korrosiv  für    Aragonit   Sauerstoffmangel  

Der  Ozean  im  globalen  Wandel  

FONA-­‐Forum,  Leipzig,  9.-­‐11.  09.  2013  

Hotspots  des    Ozeanwandels  

Page 4: Der$Ozean$im$globalen$Wandel:$ wärmer,$saurer,$atemlos$ · wärmer$ saurer$ atemlos$ sinkende$Produk]vität steigende$Produk]vität Gruber2011 korrosivfür$ Aragonit Sauerstoffmangel$

Kalifornien-­‐Strom  

Humboldt-­‐Strom  

Kanaren-­‐Strom  

Benguela-­‐Strom  

Somali-­‐Strom  

•  Kongruenz  der  globalen  Stressoren  (wärmer,  saurer,  atemlos)  •  hohe  Produk]vität    •  Schlüsselregionen  für  biogeochemische  Stoffumsätze  •  starke  Klimarückkopplung  •  hohe  ökonomische  Relevanz  à  30%  des  globalen  Fischfangs    

Aucriebssysteme:  Hotspots  im  globalen  Wandel  

FONA-­‐Forum,  Leipzig,  9.-­‐11.  09.  2013  

Page 5: Der$Ozean$im$globalen$Wandel:$ wärmer,$saurer,$atemlos$ · wärmer$ saurer$ atemlos$ sinkende$Produk]vität steigende$Produk]vität Gruber2011 korrosivfür$ Aragonit Sauerstoffmangel$

CO2 E

miss

ionen

(Gt C

/Jahr

)

Fossile Energieträger

Veränderte Landnutzung

10

8

6

4

2

1960 2010 1970 1990 2000 1980 Jahr 1 Gt = 1 Gigatonne = 1 Milliarde Tonnen

CO2  Emissionen  (1960-­‐2009)  

Global Carbon Project 2010

FONA-­‐Forum,  Leipzig,  9.-­‐11.  09.  2013  

Ursache:  anthropogene  CO2  Emissionen  

Global Carbon Project 2012

Page 6: Der$Ozean$im$globalen$Wandel:$ wärmer,$saurer,$atemlos$ · wärmer$ saurer$ atemlos$ sinkende$Produk]vität steigende$Produk]vität Gruber2011 korrosivfür$ Aragonit Sauerstoffmangel$

1.1±0.7 Gt C/Jahr

+7.7±0.5 Gt C/Jahr

2.4 Gt C/Jahr

27%

4.1±0.1 Gt C/Jahr

47%

26% 2.3±0.4 Gt C/Jahr

Global Carbon Project 2012; Angaben sind Mittel für den Zeitraum 2000-2010

CO2 + H2O à HCO3- + H+

[H+] à pH

Wo  bleibt  das  zusätzliche  CO2?  Ursache:  anthropogene  CO2  Emissionen  

FONA-­‐Forum,  Leipzig,  9.-­‐11.  09.  2013  

1 Gt = 1 Gigatonne = 1 Milliarde Tonnen

Page 7: Der$Ozean$im$globalen$Wandel:$ wärmer,$saurer,$atemlos$ · wärmer$ saurer$ atemlos$ sinkende$Produk]vität steigende$Produk]vität Gruber2011 korrosivfür$ Aragonit Sauerstoffmangel$

Mixed  layer  depth  Sinking  

Grazing  

Dienstleistungen    des  Ozeans  

Nahrungsnetze  und  Stogreisläufe  

Biodiversität  Biogeographie  Ökosysteme  

Fischerei-­‐Erträge  Wirkstoffe  aus  dem  Meer  

Stoff-­‐  und  Energiekreisläufe  

Tourismus  Küstenschutz  

CO2  Speicherung  ProdukJon  von  Klimagasen  

Globaler  Wandel  

Physik  u.  Chemie    der  Ozeane  

FONA-­‐Forum,  Leipzig,  9.-­‐11.  09.  2013  

Einfluss  auf    Organismen  

 (nega]v  und  

posi]v)  

Page 8: Der$Ozean$im$globalen$Wandel:$ wärmer,$saurer,$atemlos$ · wärmer$ saurer$ atemlos$ sinkende$Produk]vität steigende$Produk]vität Gruber2011 korrosivfür$ Aragonit Sauerstoffmangel$

Mixed  layer  depth  Sinking  

Grazing  

Dienstleistungen    des  Ozeans  

Nahrungsnetze  und  Stogreisläufe  

Biodiversität  Biogeographie  Ökosysteme  

Fischerei-­‐Erträge  Wirkstoffe  aus  dem  Meer  

Stoff-­‐  und  Energiekreisläufe  

Tourismus  Küstenschutz  

CO2  Speicherung  ProdukJon  von  Klimagasen  

Globaler  Wandel  

Physik  u.  Chemie    der  Ozeane  

FONA-­‐Forum,  Leipzig,  9.-­‐11.  09.  2013  

Einfluss  auf    Organismen  

 (nega]v  und  

posi]v)  

Page 9: Der$Ozean$im$globalen$Wandel:$ wärmer,$saurer,$atemlos$ · wärmer$ saurer$ atemlos$ sinkende$Produk]vität steigende$Produk]vität Gruber2011 korrosivfür$ Aragonit Sauerstoffmangel$

•  Wechselwirkung  mul]pler  Stressoren      global:        Erwärmung,  Versauerung,  O2  Abnahme    regional:    Eutrophierung,  Verschmutzung,  Überfischung  

•  Potenzial  zur  evolu]onären  Anpassung  

•  Sensi]vitäten  auf  Ökosystem-­‐Ebene  

•  Rückkopplungen  zum  Klimasystem  

•  Sozio-­‐ökonomische  Konsequenzen  

Forschungsdefizite  

FONA-­‐Forum,  Leipzig,  9.-­‐11.  09.  2013  

Selek]on  

Page 10: Der$Ozean$im$globalen$Wandel:$ wärmer,$saurer,$atemlos$ · wärmer$ saurer$ atemlos$ sinkende$Produk]vität steigende$Produk]vität Gruber2011 korrosivfür$ Aragonit Sauerstoffmangel$

•  starke  Veränderungen  der  Umweltbedingungen        (Kongruenz  mul]pler  Stressoren)  

•  hohe  Sensi]vität  der  Ökosysteme  •  starke  Rückkopplung  zum  Klimasystem  •  hohe  sozio-­‐ökonomische  Relevanz  

Ø  Aucriebsgebiete    Atlan]k:  Kanaren-­‐  und  Benguela-­‐Strom    Pazifik:  Humboldt-­‐  und  Kalifornien-­‐Strom    Indik:  Somali-­‐Strom,  Arabisches  Meer  

Ø  Aunau  auf  laufende  Projekte    SPACES*  (GENUS,  SACUS,  AGULHAS),  CARIMA    Ocean  Science  Centre  Mindelo,  Kapverden  *SPACES  -­‐  Science  Partnership  for  the  Assessment  of  Complex  Earth  System  Processes  

Kriterien  für  Schlüsselregionen  

FONA-­‐Forum,  Leipzig,  9.-­‐11.  09.  2013  

(Fig. 1). Equatorward winds along the eastern flanks (Figs. 1 and 2)feed the trades and drive the broad and slow eastern boundaryBenguela, California, Iberia/Canary and Humboldt currents. Nearshore (order 25–150 km), interaction with earth’s rotation (Coriolis)and presence of the coastal boundary, produces a shallow (order50 m) wind-driven offshore surface Ekman flow which is replacedby cool and nutrient-rich waters from below; such ‘coastal upwell-ing’ leaves a strong imprint on sea surface temperature and chloro-phyll (Figs. 1 and 2). Over the shelf and slope an undercurrentflows poleward so that average currents within 100 km from thecoast are opposite the surface winds; further offshore the Bengu-ela, California, Iberia/Canary and Humboldt currents flow. Thephysical process of coastal upwelling with its equations has been

described repeatedly (see Allen, 1973; Barber and Smith, 1981a;Smith, 1995; Bakun, 1996; Hill et al., 1998). In addition to thestrength of the upwelling-favorable equatorward winds, water col-umn stratification, coastal topography, and latitude-dependence ofthe Coriolis parameter play an important role; at low latitude,favorable winds produce more upwelling than at high latitude(Table 1). The shallow offshore flow, often termed Ekmantransport, is balanced by onshore subsurface flow beneath the‘Ekman layer’. Planktonic organisms, including larval fish andinvertebrates, can exploit this conveyor belt and be retained inthe coastal productive habitat (Peterson et al., 1979; Barber andSmith, 1981b; Peterson, 1998; Carr et al., 2008). The upwelling-favorable winds typically increase offshore to maximum mean

Fig. 2. Higher resolution maps of SST, winds and chlorophyll for each EBUE. SST is MODIS monthly averaged from July 2002 to April 2008; chlorophyll is SeaWiFS monthlyaveraged from September 1997 to September 2007; winds are QuikSCAT weekly averaged from July 1999 to July 2008. All products were regridded on the QuikSCAT grid(0.25! ! 0.25!) prior to averaging. The insert in the chlorophyll map shows the mean primary productivity (PP) and fish catch for the years 1998–2005. It was assumed thatthe reported fish catches (Fish and Agriculture Organization, FAO) were made within 100 km from the coast. Primary productivity was estimated from satellite remotesensing of chlorophyll and the Behrenfeld and Falkowski (1997) model. Peru fish catch exceeds that from the other areas by an order of magnitude even though PP levels aresimilar.

Table 1Comparison of annual mean properties in a 10! latitude coastal (0–150 km) band for the four EBUE. Winds are from QuikSCAT (0.25! ! 0.25!, 8-day, 1999–2008, provided by PFEL,NOAA). Calculations for upwelling, mean nitrate concentration at 60 m and the potential new production are described in Messié et al. (in this issue). Vertical upwelling rateswere calculated by dividing Ekman transport by the Rossby radius of deformation (obtained from Chelton et al., 1998) and for pumping by dividing by the 150 km band;turbulence is given as the cube of the wind speed. Chlorophyll concentration, 1 mg m"3 chl distance from shore, and PAR are calculated from SeaWiFS (9-km degraded on0.25! ! 0.25!, geometric mean); primary production was calculated following Behrenfeld and Falkowski (1997). Mixed layer depth were calculated from in situ temperatureprofiles following the Naval Research Laboratory Mixed Layer Depth (NMLD) methodology (depth where temperature = temperature at 10 m " 0.8 !C). Dust deposition wascalculated from the model of Mahowald and Luo (2003).

Benguela California NW Africa Peru28!S–18!S 34!N–44!N 12!N–22!N 16!S–6!S

Coriolis parameter f (s"1) "5.7e"05 9.18e"05 4.26e"05 "2.78e"05Area (1011 m2) 1.67Average wind speed (m s"1) 7.2 7.8 6.8 5.7Upwelled volume (Sv) 1.5 1.0 1.4 1.6Average vertical speed by ektrans (10"5 m s"1) 3.18 3.30 2.35 1.55Average vertical speed by ekpump (10"5 m s"1) 0.32 0.18 0.36 0.52Percentage ektrans/total 77.0 77.5 74.9 68.6Average [NO3] at 60 m (lmol L"l) 16.9 14.9 19.0 16.8Potential new production (g C m"2 yr"1) 517 323 539 566SeaWiFS primary production (g C m"2 yr"1) 976 479 1213 855SeaWiFS chlorophyll (mg m"3) 3.1 1.5 4.3 2.41 mg m"3 Chl extension (km) 160 49 164 120Turbulence (m3 s"3) 444 610 371 225MLD (m) 44.8 43.3 28.9 30.7PAR (E m"2 d"1) 43.1 34.0 47.6 43.3Dust deposition (g m"2 yr"1) 11.7 0.3 33.1 0.3

82 F.P. Chavez, M. Messié / Progress in Oceanography 83 (2009) 80–96

(Fig. 1). Equatorward winds along the eastern flanks (Figs. 1 and 2)feed the trades and drive the broad and slow eastern boundaryBenguela, California, Iberia/Canary and Humboldt currents. Nearshore (order 25–150 km), interaction with earth’s rotation (Coriolis)and presence of the coastal boundary, produces a shallow (order50 m) wind-driven offshore surface Ekman flow which is replacedby cool and nutrient-rich waters from below; such ‘coastal upwell-ing’ leaves a strong imprint on sea surface temperature and chloro-phyll (Figs. 1 and 2). Over the shelf and slope an undercurrentflows poleward so that average currents within 100 km from thecoast are opposite the surface winds; further offshore the Bengu-ela, California, Iberia/Canary and Humboldt currents flow. Thephysical process of coastal upwelling with its equations has been

described repeatedly (see Allen, 1973; Barber and Smith, 1981a;Smith, 1995; Bakun, 1996; Hill et al., 1998). In addition to thestrength of the upwelling-favorable equatorward winds, water col-umn stratification, coastal topography, and latitude-dependence ofthe Coriolis parameter play an important role; at low latitude,favorable winds produce more upwelling than at high latitude(Table 1). The shallow offshore flow, often termed Ekmantransport, is balanced by onshore subsurface flow beneath the‘Ekman layer’. Planktonic organisms, including larval fish andinvertebrates, can exploit this conveyor belt and be retained inthe coastal productive habitat (Peterson et al., 1979; Barber andSmith, 1981b; Peterson, 1998; Carr et al., 2008). The upwelling-favorable winds typically increase offshore to maximum mean

Fig. 2. Higher resolution maps of SST, winds and chlorophyll for each EBUE. SST is MODIS monthly averaged from July 2002 to April 2008; chlorophyll is SeaWiFS monthlyaveraged from September 1997 to September 2007; winds are QuikSCAT weekly averaged from July 1999 to July 2008. All products were regridded on the QuikSCAT grid(0.25! ! 0.25!) prior to averaging. The insert in the chlorophyll map shows the mean primary productivity (PP) and fish catch for the years 1998–2005. It was assumed thatthe reported fish catches (Fish and Agriculture Organization, FAO) were made within 100 km from the coast. Primary productivity was estimated from satellite remotesensing of chlorophyll and the Behrenfeld and Falkowski (1997) model. Peru fish catch exceeds that from the other areas by an order of magnitude even though PP levels aresimilar.

Table 1Comparison of annual mean properties in a 10! latitude coastal (0–150 km) band for the four EBUE. Winds are from QuikSCAT (0.25! ! 0.25!, 8-day, 1999–2008, provided by PFEL,NOAA). Calculations for upwelling, mean nitrate concentration at 60 m and the potential new production are described in Messié et al. (in this issue). Vertical upwelling rateswere calculated by dividing Ekman transport by the Rossby radius of deformation (obtained from Chelton et al., 1998) and for pumping by dividing by the 150 km band;turbulence is given as the cube of the wind speed. Chlorophyll concentration, 1 mg m"3 chl distance from shore, and PAR are calculated from SeaWiFS (9-km degraded on0.25! ! 0.25!, geometric mean); primary production was calculated following Behrenfeld and Falkowski (1997). Mixed layer depth were calculated from in situ temperatureprofiles following the Naval Research Laboratory Mixed Layer Depth (NMLD) methodology (depth where temperature = temperature at 10 m " 0.8 !C). Dust deposition wascalculated from the model of Mahowald and Luo (2003).

Benguela California NW Africa Peru28!S–18!S 34!N–44!N 12!N–22!N 16!S–6!S

Coriolis parameter f (s"1) "5.7e"05 9.18e"05 4.26e"05 "2.78e"05Area (1011 m2) 1.67Average wind speed (m s"1) 7.2 7.8 6.8 5.7Upwelled volume (Sv) 1.5 1.0 1.4 1.6Average vertical speed by ektrans (10"5 m s"1) 3.18 3.30 2.35 1.55Average vertical speed by ekpump (10"5 m s"1) 0.32 0.18 0.36 0.52Percentage ektrans/total 77.0 77.5 74.9 68.6Average [NO3] at 60 m (lmol L"l) 16.9 14.9 19.0 16.8Potential new production (g C m"2 yr"1) 517 323 539 566SeaWiFS primary production (g C m"2 yr"1) 976 479 1213 855SeaWiFS chlorophyll (mg m"3) 3.1 1.5 4.3 2.41 mg m"3 Chl extension (km) 160 49 164 120Turbulence (m3 s"3) 444 610 371 225MLD (m) 44.8 43.3 28.9 30.7PAR (E m"2 d"1) 43.1 34.0 47.6 43.3Dust deposition (g m"2 yr"1) 11.7 0.3 33.1 0.3

82 F.P. Chavez, M. Messié / Progress in Oceanography 83 (2009) 80–96

Primärproduk]on  

Fischerei                      Erträge  x  100  (  g

 C  /  m

2  /d  )  

Kanaren  

Humboldt  

(Fig. 1). Equatorward winds along the eastern flanks (Figs. 1 and 2)feed the trades and drive the broad and slow eastern boundaryBenguela, California, Iberia/Canary and Humboldt currents. Nearshore (order 25–150 km), interaction with earth’s rotation (Coriolis)and presence of the coastal boundary, produces a shallow (order50 m) wind-driven offshore surface Ekman flow which is replacedby cool and nutrient-rich waters from below; such ‘coastal upwell-ing’ leaves a strong imprint on sea surface temperature and chloro-phyll (Figs. 1 and 2). Over the shelf and slope an undercurrentflows poleward so that average currents within 100 km from thecoast are opposite the surface winds; further offshore the Bengu-ela, California, Iberia/Canary and Humboldt currents flow. Thephysical process of coastal upwelling with its equations has been

described repeatedly (see Allen, 1973; Barber and Smith, 1981a;Smith, 1995; Bakun, 1996; Hill et al., 1998). In addition to thestrength of the upwelling-favorable equatorward winds, water col-umn stratification, coastal topography, and latitude-dependence ofthe Coriolis parameter play an important role; at low latitude,favorable winds produce more upwelling than at high latitude(Table 1). The shallow offshore flow, often termed Ekmantransport, is balanced by onshore subsurface flow beneath the‘Ekman layer’. Planktonic organisms, including larval fish andinvertebrates, can exploit this conveyor belt and be retained inthe coastal productive habitat (Peterson et al., 1979; Barber andSmith, 1981b; Peterson, 1998; Carr et al., 2008). The upwelling-favorable winds typically increase offshore to maximum mean

Fig. 2. Higher resolution maps of SST, winds and chlorophyll for each EBUE. SST is MODIS monthly averaged from July 2002 to April 2008; chlorophyll is SeaWiFS monthlyaveraged from September 1997 to September 2007; winds are QuikSCAT weekly averaged from July 1999 to July 2008. All products were regridded on the QuikSCAT grid(0.25! ! 0.25!) prior to averaging. The insert in the chlorophyll map shows the mean primary productivity (PP) and fish catch for the years 1998–2005. It was assumed thatthe reported fish catches (Fish and Agriculture Organization, FAO) were made within 100 km from the coast. Primary productivity was estimated from satellite remotesensing of chlorophyll and the Behrenfeld and Falkowski (1997) model. Peru fish catch exceeds that from the other areas by an order of magnitude even though PP levels aresimilar.

Table 1Comparison of annual mean properties in a 10! latitude coastal (0–150 km) band for the four EBUE. Winds are from QuikSCAT (0.25! ! 0.25!, 8-day, 1999–2008, provided by PFEL,NOAA). Calculations for upwelling, mean nitrate concentration at 60 m and the potential new production are described in Messié et al. (in this issue). Vertical upwelling rateswere calculated by dividing Ekman transport by the Rossby radius of deformation (obtained from Chelton et al., 1998) and for pumping by dividing by the 150 km band;turbulence is given as the cube of the wind speed. Chlorophyll concentration, 1 mg m"3 chl distance from shore, and PAR are calculated from SeaWiFS (9-km degraded on0.25! ! 0.25!, geometric mean); primary production was calculated following Behrenfeld and Falkowski (1997). Mixed layer depth were calculated from in situ temperatureprofiles following the Naval Research Laboratory Mixed Layer Depth (NMLD) methodology (depth where temperature = temperature at 10 m " 0.8 !C). Dust deposition wascalculated from the model of Mahowald and Luo (2003).

Benguela California NW Africa Peru28!S–18!S 34!N–44!N 12!N–22!N 16!S–6!S

Coriolis parameter f (s"1) "5.7e"05 9.18e"05 4.26e"05 "2.78e"05Area (1011 m2) 1.67Average wind speed (m s"1) 7.2 7.8 6.8 5.7Upwelled volume (Sv) 1.5 1.0 1.4 1.6Average vertical speed by ektrans (10"5 m s"1) 3.18 3.30 2.35 1.55Average vertical speed by ekpump (10"5 m s"1) 0.32 0.18 0.36 0.52Percentage ektrans/total 77.0 77.5 74.9 68.6Average [NO3] at 60 m (lmol L"l) 16.9 14.9 19.0 16.8Potential new production (g C m"2 yr"1) 517 323 539 566SeaWiFS primary production (g C m"2 yr"1) 976 479 1213 855SeaWiFS chlorophyll (mg m"3) 3.1 1.5 4.3 2.41 mg m"3 Chl extension (km) 160 49 164 120Turbulence (m3 s"3) 444 610 371 225MLD (m) 44.8 43.3 28.9 30.7PAR (E m"2 d"1) 43.1 34.0 47.6 43.3Dust deposition (g m"2 yr"1) 11.7 0.3 33.1 0.3

82 F.P. Chavez, M. Messié / Progress in Oceanography 83 (2009) 80–96

(Chlorop

hyll    con

centra]o

n  (m

g    m

-­‐3)    

Benguela  

Humboldt  

Page 11: Der$Ozean$im$globalen$Wandel:$ wärmer,$saurer,$atemlos$ · wärmer$ saurer$ atemlos$ sinkende$Produk]vität steigende$Produk]vität Gruber2011 korrosivfür$ Aragonit Sauerstoffmangel$

 •  globales  Monitoring-­‐Netzwerk  für  Ozeanwandel    

 (interna]onales  Netzwerk  im  Aunau)  

•  systemare  Studien  in  Hotspot-­‐Regionen      (Beobachtungen  –  Experimente  -­‐  Modellierung)  

•  enge  Verzahnung  mit  sozio-­‐ökonomischen  Studien    (Fischerei,  Tourismus,  Küstenschutz,  ...)  

•  Kommunika]on                              (Einbindung  von  Interessensgruppen,  Informa]onsfluss                an  Entscheidungsträger  und  Öffentlichkeit)  

•  Koopera]onsbeziehungen  mit  Anrainerstaaten      (inklusive  capacity  building)  

FONA-­‐Forum,  Leipzig,  9.-­‐11.  09.  2013  

Komponenten  eines  Forschungsverbunds  

996 M. Steinacher et al.: Projected decrease in marine productivity

Fig. 10. (a) Multi-model mean of vertically integrated annual mean PP under preindustrial conditions (decadal mean 1860–1869) and(b) projected changes by the end of the 21st century under SRES A2. The changes are shown on an exponential scale and represent thedifference between 2090–2099 and 1860–1869 (decadal means). The multi-model means have been computed by using the regional skillscores shown in Fig. 9 as weights. The dotted areas indicate that none of the regional skill scores is higher than 0.5. Where no observation-based data is available to calculate skill scores (e.g. in the Arctic) the arithmetic mean of the model results is shown.

The weights w(x,y)i,j used here are the same as givenabove. The multi-model mean with this second metric is cal-culated as

PPEi,j =

m

E�1m,i,j�

mE�1m,i,j

PPm,i,j . (7)

In addition, we have computed the arithmetic mean fromall models (PPave) as well as the mean obtained by weight-ing individual models with their global (� = ⇤) skill score(PPSglob ).Next, global skill scores (Sglob) and global root mean

square errors (RMSE) are computed for the individual modelresults and for the multi-model fields obtained by the fourdifferent averaging methods (Table 1). The global skill scorefor the first field (PPS) is considerably higher than for theothers. All averaging methods result in a lower global skillscore than that of the two best models (IPSL and CCSM3).However, the RMSE is lower for the PPS field than for eachindividual model and for the other multi-model fields. In thefollowing, we discuss results from this metric only. We notethat differences in the results obtained by the first two metrics(PPS and PPE) are generally small.This skill score method accounts for the different skills

of the models at reproducing regional features of the satel-lite based estimates, while not degrading the overall skill inrepresenting the satellite-based field compared to the best in-dividual model. For example, the CSM1.4 model reproducesthe high PP tongue around 40⇥ N in the North Atlantic. TheIPSL model captures most of the high PP features along thecoasts of South America and Africa. The MPIM model hasa high skill in the central Pacific and the most realistic latitu-dinal extension of the equatorial PP belt, while the CCSM3

captures best the magnitude and pattern of PP around 40⇥ S.Therefore these models dominate the mean in those regions(Fig. 9d), and all these features are present in the multi-modelmean (Fig. 10a). There remain weaknesses. All models un-derestimate PP in the Arabian Sea and off the west coast ofNorth America. Consequently, the multi-model mean alsomisses these features. Overall, this method improves themulti-model mean significantly compared to simpler averag-ing methods (Table 1).Regional skill scores are applied to calculate the multi-

model mean of preindustrial PP and of the projected changesby the end of the 21st century (Fig. 10) and as a function ofthe global mean surface air temperature (SATglob, Fig. 11d).The globally integrated annual mean PP decreases from37.1GtC yr�1 (preindustrial) to 33.0Gt C yr�1 by 2100 AD(�2.9Gt C yr�1; �8%) for the multi-model mean (Fig. 1,Table 1). Large decreases in PP are projected for the NorthAtlantic, off the coast of Africa in the South Atlantic, in thePacific around the equator and around 30⇥, and in the north-ern part of the Indian Ocean; a slight increase in PP is foundin the Southern Ocean and in the Arctic (Fig. 10b). Calculat-ing the mean by 2100 has the disadvantage that PP changesare merged that correspond to different temperature changesas the models have different climate sensitivities. One way toavoid this is to calculate the regression slope ⌅PP/⌅SATglobfor each grid cell (Fig. 11a–c) as done for the global PP inFig. 1c. The patterns of the resulting PP change per centi-grade SAT increase are broadly consistent with the patternsof the projected PP change by 2100.

Biogeosciences, 7, 979–1005, 2010 www.biogeosciences.net/7/979/2010/

Page 12: Der$Ozean$im$globalen$Wandel:$ wärmer,$saurer,$atemlos$ · wärmer$ saurer$ atemlos$ sinkende$Produk]vität steigende$Produk]vität Gruber2011 korrosivfür$ Aragonit Sauerstoffmangel$

         

FONA-­‐Forum,  Leipzig,  9.-­‐11.  09.  2013  

Weitere  Informa]onen  

Vielen Dank für Ihr Interesse