Greening of Arctic: Knowledge and Uncertainties Jiong Jia, Hesong Wang Chinese Academy of Science jiong@tea.ac.cn Howie Epstein Skip Walker Moscow, January 28, 2008
Global Warming and Its Impact IMPACTS FOR HIGH WARMING SCENARIO OBSERVED Some increase in extreme climate events Small positive or negative net monetary impacts (most people adversely affected) Net negative for some regions Risks to some systems Risks of large scale discontinuities Net negative monetary impacts Net negative for many regions Risks to many systems Extreme and irreversible effects Aggregate impacts Distribution of impacts Unique and threatened systems Temperature anomaly ( o C) 6 5 4 3 2 1 Global mean annual temperature change relative to preindustrial IPCC High IPCC Low 0 1900 1950 2000 2050 2100 OBSERVED Some increase in extreme climate events Small positive or negative net monetary impacts (most people adversely affected) Net negative for some regions Risks to some systems IMPACTS FOR LOW WARMING SCENARIO Extreme and irreversible effects Aggregate impacts Distribution of impacts Unique and threatened systems
Available Data NASA GIMMS bi-monthly 8km 1981-2006 Recent update for 2004-2006 (September 2007) NASA PAL monthly 4km 1982-1999 (Piao( Piao,, Zhou, et al.) Long-term data record (LTDR( LTDR), daily ~2000 (Missing data over 70 degree north) MODIS 2000-2007, collection 5, 4 (ORNL subsets) Landsat MSS, TM, ETM+ over many locations since early 1970s with 5-20 yr interval (Bruce, Jiong) Near future: NPOESS-VIIRS
Knowledge Base What are the Consequences of Land Cover & Land Use Change for the Sustainability of Ecosystems & Economic Productivity? NASA Joint Unfunded Landsats-4,5 &7 Landsat-7 Landsat-7/ASTER IKONOS Data buy Aqua/MODIS Terra/MODIS, MISR & ASTER NOAA AVHRR Aqua/MODIS AVHRR global 1 km land cover; Local to regional scale inventories of land cover & land use change using Landsat 5 data. Terra/ MODIS Spatially explicit socio-economic & ecological models of land cover/ land use change working synergistically NPP-VIIRS Landsat Data Continuity Mission NPOESS- VIIRS Land Use & Integrated Assessment Models Ecological, biogeochemical, and land use model simulation results incorporating actual land cover observations Quantitative inventory of global forest cover Understanding the physical and socio-economic impacts of land use changes Global and regional land cover change products detailing loss of biodiversity, habitat degradation and fragmentation, changes in water resources and changes in the forestry and agriculture sectors; changes in urban areas, wetlands and coastal zones Global estimation of above-ground biomass carbon storage and sources Land cover and land use change effects on atmospheric, surface radiation and hydrological processes Environmental & Climate Data Records at moderate spatial resolution Continuous high-resolution global seasonally-refreshed land cover and land use data Continued improvement in algorithms for land cover classification, direct parameterizations and change detection Evaluate consequences of observed and predicted changes and further understanding of their impact on environmental goods and services, the carbon and water cycles and the management of natural resources 2002 2004 2006 2008 2010 2012 2014 2015 Forecast the extent & impacts of land use change at local, national, and regional scales with models integrating imagery across scales (e.g., MODIS to IKONOS) with socioeconomic & climate data
AVHRR Data Corrections Compared to newer sensors (e.g. MODIS, MISR), AVHRR lacks of onboard calibration; Low SNR due to cloud contamination and water vapor (Mostly corrected with MVC); Geometric registration errors (Mostly corrected); Volcanic aerosol effects for 1982-1984 and 1991-1994 (partly corrected); Residual sensor degradation and viewing angle effects due to satellite drift (partly corrected, problem over recent years)
Use of EMD technique and the removal of SZAcorrelated trends Top: Before draft correction, Middle: trend removed Bottom: resulting series.
What we have done Unmix the signal of vegetation and select areas with relatively homogeneous vegetation with MODIS and Landsat ETM+ for each subzone and boreal forest; Mask pixels of water and ice/snow in analysis; Temporal-spatial analysis of peak and time-integrated NDVI by continents and subzones Analysis of vegetation phenology Examining the uncertainties in interpreting trends of greenness
Observations Remotely sensed In situ measurements Meta analysis Scaling up
Modeled Landscape
Circumpolar Greenness
Vegetation in Yamal
Vegetation shown on photos and Landsat ETM+ images Subzone A Subzone B Subzone C
Vegetation shown on photos and Landsat ETM+ images Subzone D Subzone E
1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Circumpolar Peak Vegetation Greenness 1981-2003
1981-1986 1987-1992 Semi-decadal comparison of peak NDVI over circumpolar region 1993-1998 1999-2003
Annual peak NDVI 0.6 0.55 0.5 0.45 Yamal Arctic tundra +0.44%/yr, R2=0.61, p<0.001 0.4 1981 1986 1991 Years 1996 2001 2006 0.8! 0.75 Yamal Arctic south of 70 degree north +0.34%/yr, r2=0.36, p<0.01 Changes of annual peak vegetation greenness (NDVI) over tundra biome from 1982-2003 (top) and below 70 degree north from 1982-2005 (bottom) in Yamal Arctic, Russia as detected by NOAA AVHRR time series data. Annual peak NDVI 0.7 0.65 0.6 0.55 0.5 1981 1986 1991 1996 2001 2006 Years
0.60 Annual peak NDVI 0.55 0.50 0.45 North American Arctic 0.64%/yr, r2=0.71, p<0.001 0.40 1981 1986 1991 1996 2001 2006 Years 0.8 Changes of annual peak vegetation greenness (NDVI) over tundra biome from 1982-2003 (top) and below 70 degree north from 1982-2005 (bottom) in North American Arctic (Alaska and Canada) as detected by NOAA AVHRR time series data. Annual peak NDVI 0.75 0.7 0.65 0.6 0.55 North American Arctic south of 70 degree 0.58%/yr, r2=0.70, p<0.001 0.5 1981 1986 1991 1996 2001 2006 Years
Canadian High Arctic 0.35 0.3 Peak NDVI 0.25 0.2 0.15 0.1 0.05 Subzone A: 0.492%/yr Subzone B: 0.55%/yr Subzone C: 0.806%/yr 82 84 86 88 90 92 94 96 98 00 02 Year
Canadian Low Arctic 0.9 0.8 Peak NDVI 0.7 0.6 0.5 0.4 0.3 Subzone D: 0.672%/yr Subzone E: 0.354%/yr Taiga: -0.067%/yr 82 84 86 88 90 92 94 96 98 00 02 Year
Russia High Arctic 0.6 0.5 Peak NDVI 0.4 0.3 0.2 0.1 0.0 Subzone A: 0.556%/yr Subzone B: 0.537%/yr Subzone C: 0.407%/yr 81 83 85 87 89 91 93 95 97 99 01 03 Year
Russia Low Arctic and Taiga 1.00 0.90 Peak NDVI 0.80 0.70 0.60 0.50 0.40 Subzone D: 0.371%/yr Subzone E: 0.283%/yr Taiga: -0.073%/yr 81 83 85 87 89 91 93 95 97 99 01 03 Year
Seasonal Pattern of Vegetation Greenness
NDVI 0.18 0.17 0.16 0.15 0.14 0.13 0.12 0.11 Subzone A 1982-1992 Subzone A 1993-2003 A Arctic Subzone A 0.10 5a 5b 6a 6b 7a 7b 8a 8b 9a 9b 0.22 0.20 Subzone B 1982-1992 Subzone B 1993-2003 B 0.18 NDVI 0.16 Arctic Subzone B 0.14 0.12 0.10 5a 5b 6a 6b 7a 7b 8a 8b 9a 9b
0.30 0.28 0.26 0.24 Subzone C 1982-1992 Subzone C 1993-2003 C NDVI 0.22 0.20 0.18 0.16 Arctic Subzone C 0.14 0.12 0.10 5a 5b 6a 6b 7a 7b 8a 8b 9a 9b 0.60 0.55 0.50 0.45 Subzone D 1982-1992 Subzone D 1993-2003 D Arctic Subzone D NDVI 0.40 0.35 0.30 0.25 0.20 0.15 0.10 5a 5b 6a 6b 7a 7b 8a 8b 9a 9b
0.70 0.60 Subzone E 1982-1992 Subzone E 1993-2003 0.50 E NDVI 0.40 0.30 Arctic Subzone E 0.20 0.10 5a 5b 6a 6b 7a 7b 8a 8b 9a 9b Bimonthly periods
What s s Behind the Greening General trend of global and regional warming Stronger warming signals Get closer to Canadian Arctic (1km ~2005) Revisit Arctic Alaska (1km, weekly, ~2006) Spatial patterns of NDVI changes Impact of surface temperature and sea-ice on vegetation greenness
MODIS + AVHRR over Siberia 1982-2005 (Graversen et al, 2008, Nature)
Temperature anomalies (Graversen et al, 2008, Nature)
Fire and NPP anomalies (Bond-Lamberty, 2007, Nature)
Accumulative effect on biomass and NPP Standing biomass Cumulative NPP Time Time Exclosure Low/Moderate intensity grazing High intensity grazing
Uncertainties Major uncertainties came from data calibrations; Draft of AVHRR sensors and differences in spectra and algorisms between AVHRR and MODIS; Confused from pixels contain mixture of land cover types respond to warming in different ways; Time periods considered in analysis; Geographic extents in analysis; Reluctance in facing the inconvenient truth
MODIS vs. AVHRR NDVI
AVHRR over Siberia 1982-2003
MODIS + AVHRR over Siberia 1982-2005
0.5 Peak NDVI 0.4 0.3 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 Year 0.7 Peak NDVI 0.6 0.5 0.4 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 Year How trend of greening is affected by the length and period of time series
3.0! TI-NDVI TI-NDVI 2.5 2.0 1.5 1.0 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 Year! 4.5 4.0 3.5 3.0 2.5 2.0 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 Year In case of time-integrated NDVI
Goetz et al. (2005)
80966 80666 81966 81666 8:966 8:666 89966 +%,%-./.012%3%41.-. 5. %,%6/6780 GAC Peak NDVI NA boreal 89666 8706 8709 8776 8779.666.669.686 680-- 68--- 670-- GAC summer NDVI NA boreal 67--- 630-- 63--- 600-- 60--- *%+%,-.,/01%2%3434/ 5 / %+%-.6,63 6,0-- 6,--- 698-6980 699-6990 /--- /--0 /-6-
7<000 7/000 7;000 7:000 72000 73000 77000 -&.&/012345&6&70/748 9 3 &.&0173:7 GAC Peak NDVI NA Subzone D 70000 7480 748; 7440 744; 3000 300; 3070 372:: 37::: 322:: GAC Peak NDVI NA Subzone E 32::: 3<2:: 3<::: -(.(/012345(6(24/77 8 9 (.(:1304; 3/2:: 34;: 34;2 344: 3442 9::: 9::2 9:3:
Plans for 2008 Revise JGR paper; Submit Alaska revisit paper; Prepare and submit global trends and uncertainty paper; Resubmit Arctic vegetation RS proposal to NSFC, with Howie and Skip int l l collaborators; Continue support a student on GOA work
Thanks! Questions?
Bunn et al, 2007, EOS