Changes in growing season duration and productivity of northern vegetation inferred from long-term remote sensing data

Abstract

Monitoring and understanding climate-induced changes in the boreal and arctic vegetation is critical
\nto aid in prognosticating their future.Weused a 33 year (1982–2014) long record of satellite
\nobservations to robustly assess changes in metrics of growing season (onset: SOS, end: EOS and length:
\nLOS) and seasonal total gross primary productivity. Particular attention was paid to evaluating the
\naccuracy of these metrics by comparing them to multiple independent direct and indirect growing
\nseason and productivity measures. These comparisons reveal that the derived metrics capture the
\nspatio-temporal variations and trends with acceptable significance level (generally p<0.05).We find
\nthat LOS has lengthened by 2.60 d dec−1 (p<0.05) due to an earlier onset of SOS (−1.61 d dec−1,
\np<0.05) and a delayed EOS (0.67 d dec−1, p<0.1) at the circumpolar scale over the past three
\ndecades. Relatively greater rates of changes in growing season were observed in Eurasia (EA) and in
\nboreal regions than in North America (NA) and the arctic regions. However, this tendency of earlier
\nSOS and delayed EOS was prominent only during the earlier part of the data record (1982–1999).
\nDuring the later part (2000–2014), this tendency was reversed, i.e. delayed SOS and earlier EOS. As for
\nseasonal total productivity, we find that 42.0% of northern vegetation shows a statistically significant
\n(p<0.1) greening trend over the last three decades. This greening translates to a 20.9% gain in
\nproductivity since 1982. In contrast, only 2.5% of northern vegetation shows browning, or a 1.2% loss
\nof productivity. These trends in productivity were continuous through the period of record, unlike
\nchanges in growing season metrics. Similarly, we find relatively greater increasing rates of productivity
\nin EA and in arctic regions than inNAand the boreal regions. These results highlight spatially and
\ntemporally varying vegetation dynamics and are reflective of biome-specific responses of northern
\nvegetation during last three decades.
\nphotosynthetically active growing season, gross primary productivity, boreal and arctic, remote sensing, climate change,
\nAVHRR, MODIS