Fig 1.
(a) Spatial pattern of the Holdridge life eco-zones in the Northern Hemisphere. (b) Distribution of vegetation types in the Northern Hemisphere (adapted from UMD Global Land Cover Classification, 1 Kilometer, Department of Geography, University of Maryland).
Fig 2.
Schematic figures illustrating method identifying the beginning of growing season (BGS), end of growing season (EGS), length of growing season (LGS) and peak of growing season (PGS).
Fig 3.
Spatial distributions of the mean (a) BGS, (b) EGS, (c) LGS and (d) PGS in the northern hemisphere during 2002–2012.
The results were calculated by 11-year MODIS NDVI data.
Fig 4.
Spatial distribution of the changes in dates of BGS, EGS, LGS and PGS over the past three decades in the Northern Hemisphere.
Positive values (red colors) represent later onset (BGS), later finish (EGS), longer duration (LGS) and later peak of the growing season. (a), (e), (i): changes in dates of vegetation green-up for 1982–1992, 1992–2002 and 2002–2012, respectively (days yr-1). (b), (f), (k): changes in dates of vegetation senescence for 1982–1992, 1992–2002 and 2002–2012, respectively (days yr-1). (c), (g), (l): changes in dates of vegetation growing season length for 1982–1992, 1992–2002 and 2002–2012, respectively (days yr-1). (d), (h), (m): changes in dates of vegetation growing peak for 1982–1992, 1992–2002 and 2002–2012, respectively (days yr-1).
Fig 5.
Overall trends of phenological transitional dates (DOY, day of year) along the longitudinal gradient in Northern Hemisphere during 1982–2012.
(a) overall trends in dates of vegetation phenology for 1982–2012. (b) variability of vegetation phenology along the latitudinal gradient from 1982 to 2002.The dark solid line indicates the mean and the shaded zone indicates the range of interannual variability.
Table 1.
Shifts of phenology metrics based on NOAA AVHRR data (1982–2002) and MODIS data (2002–2012) (days per decade).
Fig 6.
Interannual variations of (a) BGS, (b) EGS, (c) LGS and (d) PGS for specific vegetation classes over Northern Hemisphere based on AVHRR data (back lines, 1982–2006) and MODIS data (red lines, 2002–2012).
Fig 7.
Spatial distribution of the change rate of phenology shift (RPS) within different Holdridge life eco-zones during the growing season from 1982 to 2012 in the Northern Hemisphere.
Positive values (red colors) represent later onset (BGS), later finish (EGS), longer duration (LGS) of phenology in a time series. (a), (d), (g): RPS of BGS for 1982–1992, 1992–2002 and 2002–2012, respectively. (b), (e), (h): RPS of EGS for 1982–1992, 1992–2002 and 2002–2012, respectively. (c), (f), (i): RPS of LGS for 1982–1992, 1992–2002 and 2002–2012, respectively.
Fig 8.
Statistical distribution of Correlation coefficients of the interannual variations of BGS, EGS and of the climatic data (spring temperature, autumn temperature and mean annual precipitation) in the Northern Hemisphere during 1982–2012.
Fig 9.
Distribution of BGS, EGS changes in relation to the mean annual temperature (upper), and the frequency of pixel in relation to mean annual temperature (bottom) within the different Holdridge life eco-zones during 1982–2012.
We used the 30-year mean annual temperature to represent the climatic temperature condition. Only pixels that BGS and EGS related to temperature for statistically significant at 90% confidence level are included. (Holdridge life eco-zones including boreal coniferous forest (Ba), boreal tundra woodland (Bb), boreal mountain system (BM), temperate continental forest (TeDc), and temperate steppe (TeBSK)).
Fig 10.
Sensitivity of vegetation phenological changes to temperature change within the different Holdridge life eco-zones during 1982–2012.
The 30-year mean annual temperature was used to represent the climatic temperature condition. Only pixels that BGS and EGS related to temperature for statistically significant at 90% confidence level are included.