The Greening Earth *department of geography, boston university, cybele.bu.edu Ranga B. Myneni* & Compton ‘Jim’ Tucker With contributions from: Alexeyev, Anderson, Asrar, Bogaert, Bousquet, Buermann, Ceulemans, Cramer, Dickinson, Dong, Friedlingstein, Hashimoto, Hughes, Jolly, Kaufmann, Kauppi, Keeling, Knyazikhin, Lucht, Liski, Nemani, Piper, Potter, Prentice, Running, Shabanov, Sitch, Slayback, Song, Smith and Zhou This research was funded by NASA-ESE. (1 of 37) Gustav Klimt ( ): Der Park (1910) MOMA-NY

publications Myneni, R. B., et al., Increased plant growth in the northern high latitudes from Nature, 386: Myneni, R. B., et al., Interannual variations in satellite-sensed vegetation index data from 1981 to J. Geophys. Res., 103 (D6): Kaufmann et al., Effect of orbital drift and sensor changes on the time series of AVHRR vegetation index data. IEEE Trans. Geosci. Remote Sens., 38: Zhou et al., Variations in northern vegetation activity inferred from satellite data of vegetation index during 1981 to 1999, J. Geophys. Res., 106 (D17): Myneni and Dong et al., A large carbon sink in the woody biomass of northern forests. Proc. Natl. Acad. Sci. USA., 98(26): Tucker et al., Higher northern latitude NDVI and growing season trends from 1982 to Int. J. Biometeorol., 45: Shabanov et al., Analysis of interannual changes in northern vegetation activity observed in AVHRR data during 1981 to IEEE Trans. Geosci. Remote Sens., 40: Lucht et al., Climatic control of the high-latitude vegetation greening trend and Pinatubo effect. Science, 296: (May ). Bogaert et al., Evidence for a persistent and extensive greening trend in Eurasia inferred from satellite vegetation index data. J. Geophys. Res., Vol. 107 (D11), /2001JD Kaufmann et al., Reply to Comment on "Variations in northern vegetation activity inferred from satellite data of vegetation index during " by J. R. Ahlbeck. J. Geophys. Res., Vol. 107(D11), /2001JD Dong et al., Remote sensing of boreal and temperate forest woody biomass: Carbon pools, Sources and Sinks, Remote Sens. Environ. 84: 393–410. Buermann et al., Circulation anomalies explain interannual covariability in northern hemisphere temperatures and greenness. J. Geophys. Res. (accepted Dec 2002). Zhou et al., Relation between interannual variations in satellite measures of vegetation greenness and climate between 1982 and J. Geophys. Res. 108(D1), doi: /2002JD Nemani et al., Climate driven increases in terrestrial net primary production from 1982 to Science (in review Apr 2003). (2 of 37)

outline Background on NDVI and NDVI data sets - the northern latitude greening trend during the 1980s and 1990s - persistence of greening in Eurasia vs North America - the temperature connection - the connection to circulation anomalies - northern latitude greening and the forest woody biomass carbon sink The greening earth and increasing terrestrial net primary production Greening in the north (3 of 37)

reflectance spectrum of a green leaf Pigments in green leaves (notably chlorophyll) absorb strongly at red and blue wavelengths. Lack of such absorption at near-infrared wavelengths results in strong scatter from leaves. upper epidermis palisade layer spongy tissue lower epidermis image credit: Govaerts et al. (4 of 37)

normalized difference vegetation index, ndvi The contrast between red and near-infrared reflectance of vegetation is captured by the greenness index, NDVI, as [(nir-red)/(nir+red)]. image credit: Huete et al. (5 of 37)

avhrr ndvi data sets The Advanced Very High Resolution Radiometers, AVHRR, have been flown on NOAA polar orbiting afternoon-viewing platforms NOAA-07: jul 81 to jan 85 NOAA-09: feb 85 to oct 88 NOAA-11: nov 88 to sep 94 NOAA-14: jan 85 to oct 01. (6 of 37) GIMMS and PAL NDVI Data Sets Have Calibration Partial Atmospheric Correction Corrections for Stratospheric Aerosols 10 or 15-day Maximum Value Composites Run from July 1981 to about mid-2001 GIMMS: GLOBAL INVENTORY MONITORING AND MODELING SYSTEMS PAL: PATHFINDER AVHRR LAND

outline Background on NDVI and NDVI data sets - the northern latitude greening trend during the 1980s and 1990s - persistence of greening in Eurasia vs North America - the temperature connection - the connection to circulation anomalies - northern latitude greening and the forest woody biomass carbon sink The greening earth and increasing terrestrial net primary production Greening in the north Use with caution - download from (7 of 37)

greening trend in the north JanDecJul Aug earlier spring delayed fall JanDecJul Aug Increase changes in growing season duration changes in greenness magnitude In the north, where vegetation growth is seasonal, the cumulative growing season greenness, which is the area under the NDVI curve, can change either due to a longer photosynthetically active growing season or due to increased greenness magnitude, or both. 1) Define vegetated pixels in the study area using a land cover map 2) Use NDVI values greater than zero only to avoid sparsely vegetated areas, pixels with snow and any corrupted data 3) Assess changes in peak seasonal greenness from July and August average NDVI 4) Use NDVI threshold to assess changes in dates of spring green-up and autumn green- down (assess sensitivity to threshold value) (8 of 37)

greening trend in the north (1980s & 90s) NDVI averaged over boreal growing season months of May to September increased by about 10%, the timing of spring green-up advanced by about 6 days. NDVI averaged over boreal growing season months of May to September increased by about 10%, the timing of spring green-up advanced by about 6 days. 8.4%/18 yrs (p<0.05) 12.4%/18 yrs (p<0.05) 11.9 days/18 yrs (p<0.05) 17.5 days/18 yrs (p<0.05) Analysis of GIMMS (v1) ndvi data for the period 1981 to 1999 indicate that: A larger increase in growing season NDVI magnitude (12 vs 8%) and a longer active growing season (18 vs 12 days) brought about by an early spring and delayed autumn are observed in Eurasia relative to North America (9 of 37) From Zhou et al., (JGR, 106(D17): , 2001)

the northern latitudes (>40N) have greened since the early 1980s outline Background on NDVI and NDVI data sets - the northern latitude greening trend during the 1980s and 1990s - persistence of greening in Eurasia vs North America - the temperature connection - the connection to circulation anomalies - northern latitude greening and the forest woody biomass carbon sink The greening earth and increasing terrestrial net primary production Greening in the north Use with caution – download from (10 of 37)

spatial pattern of greening From Zhou et al., (JGR, 106(D17): , 2001) Analyses of pixel-based persistence indices from GIMMS (v1) NDVI data for the period 1981 to 1999 indicate that: About 61% of the total vegetated area between 40N-70N in Eurasia shows a persistent increase in growing season NDVI over a broad contiguous swath of land from Central Europe through Siberia to the Aldan plateau, where almost 58% (7.3 million km2) is forests and woodlands. North America, in comparison, shows a fragmented pattern of change, notable only in the forests of the southeast and grasslands of the upper Midwest. These results are further substantiated from a study that evaluated patch characteristics using landscape ecology metrics (Bogaert et al., JGR, 107(D11): /2001jd001075, 2002). (11 of 37)

the greening in Eurasia is more persistent than in North America outline the northern latitudes (>40N) have greened since the early 1980s Background on NDVI and NDVI data sets - the northern latitude greening trend during the 1980s and 1990s - persistence of greening in Eurasia vs North America - the temperature connection - the connection to circulation anomalies - northern latitude greening and the forest woody biomass carbon sink The greening earth and increasing terrestrial net primary production Greening in the north Use with caution – download from (12 of 37)

greenness and surface temperature yx A statistically meaningful relation? y =  0 +  1 x +  y =  0 +  1 x+  2 time +   y =  0 +  1  x +  EA NDVIEA Tyes NA NDVINA Tyes EA NDVINA Tno NA NDVIEA Tno T – GISS temperature; EA – Eurasia; NA - North America From Zhou et al., (JGR, 106(D17): , 2001) The temporal changes and continental differences in NDVI are consistent with ground based measurements of temperature, an important determinant of biological activity in the north (13 of 37)

modeling the temperature connection From Lucht et al., (Science, 296: , 2002) A biogeochemical model of vegetation using observed climate data predicted the high northern latitude greening trend over the past two decades observed by satellites and a marked setback in this trend after the Mount Pinatubo volcano eruption in The observed trend toward earlier spring budburst and increased maximum leaf area is produced by the model as a consequence of biogeochemical vegetation responses mainly to changes in temperature. (14 of 37)

it is temperature, not co2, that is the related to greening outline the greening in Eurasia is more persistent than in North America the northern latitudes (>40N) have greened since the early 1980s Background on NDVI and NDVI data sets - the northern latitude greening trend during the 1980s and 1990s - persistence of greening in Eurasia vs North America - the temperature connection - the connection to circulation anomalies - northern latitude greening and the forest woody biomass carbon sink The greening earth and increasing terrestrial net primary production Greening in the north Use with caution – download from (15 of 37)

cca analysis are these relations valid at finer (pixel) scales? what is causing these correlations? NDVI data: FASIR GIMMS v1 ( ) (Courtesy of Los et al.) Surface temperature from GISS (Hansen et al., 1999) NINO3 index (Reynolds and Smith, 1994) AO index (first EOF of NH SLP 20N-90N; Thompson and Wallace, 1998) Data From Buermann and Anderson et al. (JGR, in press) Zhou et al., (2001) Canonical Correlation Analysis (CCA) to isolate coupled spatial patterns between temperature and NDVI and assess their possible relationship to large-scale circulation anomalies Method (16 of 37)

the enso mode First canonical factor: the ENSO signal Temperature (r=0.78) NDVI (r=0.76) NINO3 SON (81-97) Principal Component Spring (MAM) Temp and Spring NDVI, 10N-90N Geographic Plot of the Grid-point Correlations for the 1 st CF TEMPERATURE NDVI These figures indicate that the first factor captures the NH spring ENSO tele- connection signal in the surface temperature and NDVI fields. During warm ENSO events, warmer and greener conditions prevail in spring over North America, far east Asia and to some extent over Europe. The ENSO related patterns explain 10.8% (13.5%) of the total spring surface temperature (NDVI) variability. (17 of 37)

the ao mode Temperature (r=0.85 ) NDVI (r=0.83) AO (DJB), 82-98NAO (DJF), Principal Component Spring (MAM) Temp and Spring NDVI, 10N-90N Second canonical factor: the AO signal Geographic Plot of the Grid-point Correlations for the 2nd CF TEMPERATURE NDVI The fact that there is quantitative agreement between the temporal and spatial features isolated via the CCA algorithm and those associated with ENSO and AO indices suggests that surface temperature signatures associated with these two predominant modes of global climate variability are also important drivers for variability in northern hemisphere greenness. (18 of 37) These figures indicate that the first factor captures the NH spring AO teleconnection signal in the surface temperature and NDVI fields.

ENSO and AO are partly responsible for the correlation between temperature and NDVI outline it is temperature, not co2, that is the related to greening the greening in Eurasia is more persistent than in North America the northern latitudes (>40N) have greened since the early 1980s Background on NDVI and NDVI data sets - the northern latitude greening trend during the 1980s and 1990s - persistence of greening in Eurasia vs North America - the temperature connection - the connection to circulation anomalies - northern latitude greening and the forest woody biomass carbon sink The greening earth and increasing terrestrial net primary production Greening in the north Use with caution – download from (19 of 37)

biomass carbon stocks, sources and sinks From Myneni and Dong et al. (PNAS, 98(26): , 2001) About 1 to 2 giga (10^9) tons of carbon (Gt C) a year are suggested to be sequestered in pools on northern land. This study is limited to analysis of the carbon pool in the woody biomass of northern temperate and boreal forests, which cover an area of about 1.4 to 1.5 billion hectares. We define forests as the following remote sensing land covers: broad leaf forests, needle leaf forests, mixed forests and woody savannas. Motivation (20 of 37) Debate is currently underway regarding which of the forest biomass sinks can be used by the industrialized nations to meet their commitments under the Kyoto protocol.

inventory data and greenness Year-to-year changes in biomass are quite small, about two orders of magnitude smaller than the biomass pool. At decadal and longer time scales, the biomass changes can be considerable due to accrual of the differences between gains and losses. Potentially, these can be observed as low frequency variations in decadal scale greenness, in much the same way as century scale greenness changes are suggestive of successional changes. We use 5-yr averages of growing season NDVI total (GIMMS v1), the area under seasonal NDVI curve and above a threshold, which captures both the average seasonal level of greenness and growing season duration, and therefore is an ideal measure of seasonal greenness. Forest inventory data from 171 provinces in six countries that represent a wide variety of inventory practices, provincial forest area, ecosystem types, age structures and time periods. (21 of 37)

uncertainty analysis The relation between woody biomass and seasonal greenness is estimated as, 1/biomass = a +  [(1/ndvi)/latitude^2] +  latitude where, biomass: inventory estimate (tons/ha), ndvi: cumulative growing season ndvi averaged over five years of the inventory, latitude: average of latitudes over forest pixels in each province and, ,  and  : regression coefficients The relative difference between remote sensing and inventory estimates is 27% for above-stump biomass (10.4 tons C/ha), 33% for total biomass (16.1 tons C/ha), 50% for changes in pool size (0.33 tons C/ha/yr) (22 of 37)

spatial picture of the biomass sink Pool changes were then evaluated as the difference between the late 1990s and early 1980s pool estimates, pixel-by-pixel, and quoted on a per year basis. The carbon pool in the woody biomass of northern forests (1.5 billion ha) is estimated to be 61  20 Gt C during the late 1990s. Our sink estimate for the woody biomass during the 1980s and 1990s is 0.68  0.34 Gt C/yr. (23 of 37)

country analysis The estimates of the three large countries, Canada, Russia and the USA, are crucial because they account for 78% of the pool, 73% of the sink and 77% of the forest area. (24 of 37) For Canada, we estimate a sink of about 73 Mt C/yr which is comparable to an inventory estimate by the Canadian forest service about 85 Mt C/yr.

country analysis (25 of 37) Our pool, sink and forest area estimates for the USA are are comparable to TBFRA-2000 estimates. Our sink estimate for the USA (142 Mt C/yr) is comparable to most estimates for the 1980s (110 to 150 Mt C/yr).

country analysis (26 of 37) Estimates for Russia differ principally because of differences in the definition of forest area. When expressed on a per unit forest area basis, the various estimates are comparable.

reasons The spatial patterns, however, offer some clues: a) longer growing seasons from warming in the northern latitudes possibly explain some of the changes, and b) increased incidences of fires and infestations in Canada c) fire suppression and forest re-growth in the USA d) declining harvests in Russia e) improved silviculture in Nordic and European countries f) forest expansion and re-growth in China (27 of 37)

biomass carbon sinks represent 10% of the annual fossil fuel emissions ENSO and AO are partly responsible for the correlation between temperature and NDVI it is temperature, not co2, that is the related to greening the greening in Eurasia is more persistent than in North America the northern latitudes (>40N) have greened since the early 1980s Background on NDVI and NDVI data sets - the northern latitude greening trend during the 1980s and 1990s - persistence of greening in Eurasia vs North America - the temperature connection - the connection to circulation anomalies - northern latitude greening and the forest woody biomass carbon sink The greening earth and increasing terrestrial net primary production Greening in the north Use with caution – download from outline (28 of 37)

motivation Global environmental changes between 1980 and 2000 have been significant: - Two of the warmest decades in the instrumental record - Three intense El Nino events ( ; ; ) - Changes in tropical cloudiness and monsoon dynamics - A 9.3% increase in atmospheric co2 concentration - A 36% increase in global population (4.45 billion in 1980 to 6.08 billion in 2000) A substantial incentive to understand trends and variability in terrestrial Net Primary Production because NPP: - is the foundation of food, fiber and fuel for human consumption - determines seasonal and interannual variations in atmospheric co2 - integrates climatic, ecological, geochemical and human influences on the biosphere How have global environmental changes affected (eased or strengthened) climatic constraints to plant growth and NPP? Image credit: IPCC Image credit: FAO (29 of 37) From Nemani et al., (Science; in review, 2003)

step 1: limiting factors Plant growth is assumed to be principally limited by sub-optimal climatic conditions such as low temperatures, inadequate rainfall and cloudiness (Churkina and Running, 1998). We used average climate data (Leemans and Cramer, 1991) to develop scaling factors between 0 and 1 that indicate the reduction in growth potential. Dominant Controls water availability 40% temperature 33% solar radiation 27% Total vegetated area: 117 M km2 (30 of 37)

step 2: trends in climate data Interannual trends in daily average temperature Interannual trend in vapor pressure deficit Interannual trend in solar radiation Data: Reanalysis data (6-hourly 2 m height temperatures, 2-m height specific humidity, and incident solar radiation) from NCEP to represent climate variability from 1982 to Potential Climate Limits for Plant Growth The observed climatic changes have been mostly in the direction of reducing climatic constraints to plant growth. Therefore, it seems likely that vegetation responded to such changes positively. (31 of 37)

step 3: npp evaluation Step 1 convert absorbed radiation to optimal gross production Step 2 downgrade by climate limiting factors to obtain gpp Step 3 subtract respiration to obtain npp Average of interannual trends ( ) in growing season NPP estimated with GIMMS and PAL (v3) FPAR Trends in NPP are positive over 55% of the global vegetated area and are statistically more significant than the declining trends observed over 19% of the vegetated area. The NPP Algorithm (32 of 37)

climate, npp and atmospheric co2 growth rate Analyses of variation in the plant photosynthesis- respiration balance, expressed as NPP/GPP ratio (right panel), showed observed declines in NPP during El Niño years to be dominated by increases in respiration due to warmer temperatures. Interannual variations in global NPP are correlated with global atmospheric CO2 growth rates (r = 0.70, p<0.001). NPP declined during all three El Niño events. Although the atmospheric co2 growth rate depends on the net air-sea and land- atmosphere exchanges, these results highlight the preeminent role of plant growth in global carbon cycle. (33 of 37)

npp trends by latitude A strong decline in NPP following the Mt. Pinatubo eruption (1991) was evident only at the high latitudes of the Northern Hemisphere. Cooler temperatures resulting from the eruption decreased the growing season length at high latitude. Ecosystems in all tropical regions and those in the high latitudes of the Northern Hemisphere accounted for 80% of the increase in global NPP between 1982 and El Niño impacts are strong at low latitudes when compared to mid- and high latitudes. The same cooling may have promoted plant growth in low latitude ecosystems by reducing the evaporative demand and respiration losses. (34 of 37)

npp trends in the tropics We suggest increases in solar radiation, as a result of declining cloud cover, in these predominantly radiation-limited forests as a plausible explanation for the increased NPP. The evergreen forests of the Amazon region showed NPP increases, on average, of >1.0% / yr, contributing to over 40% of the global NPP increases between 1982 and An increase in NPP of <0.22% per ppm increase in CO2, within the range of experimental evidence, could be invoked to explain all of the estimated global NPP increase of 6.17% / 18yr. NPP increases of >1%/yr as in the case of Amazonia require a fertilization effect greater than 0.5% per ppm of CO2 increase, which appears to be much greater than those reported by field experiments. (35 of 37)

Tropical and northern ecosystems drive increases in terrestrial npp as a result of easing of climatic limits to plant growth outline biomass carbon sinks represent 10% of the annual fossil fuel emissions ENSO and AO are partly responsible for the correlation between temperature and NDVI it is temperature, not co2, that is the related to greening the greening in Eurasia is more persistent than in North America the northern latitudes (>40N) have greened since the early 1980s Background on NDVI and NDVI data sets - the northern latitude greening trend during the 1980s and 1990s - persistence of greening in Eurasia vs North America - the temperature connection - the connection to circulation anomalies - northern latitude greening and the forest woody biomass carbon sink The greening earth and increasing terrestrial net primary production Greening in the north Use with caution – download from (36 of 37)

bottomline - half the vegetated lands greened by about 11% - 15% of the vegetated lands browned by about 3% - 1/3 rd of the vegetated lands showed no changes. Since the early 1980s about, The End (37 of 37) These changes are due to easing of climatic constraints to plant growth.