Search for a command to run...
NOAA's recent assessment of Arctic greenness has reported a remarkable finding: the Arctic is browning (Epstein et al., 2015). Whilst a clear greening trend has been apparent for most of the satellite record's 33 year history (indicating an increase in biomass and productivity), there is now an overall decline in greenness from 2011 to 2014. If this is a new direction of travel for arctic vegetation, rather than just a temporary departure from long-term greening, this has major implications for not only our understanding of the future of arctic vegetation, but also arctic carbon, nutrient and water cycling, surface energy balance and permafrost degradation, and therefore feedback to climate, all of which are strongly influenced by vegetation composition, productivity and biomass. The urgency in understanding what is happening here is clear. Most models predict arctic greening; to what extent are they wrong, and why? Arctic greening has rightly received much attention. Satellite and observational data have consistently confirmed an increase in vegetation cover and productivity in many regions (Xu et al., 2013) caused most notably by the expansion of large stature deciduous shrubs (Myers-Smith et al., 2015). Likewise, field simulation experiments provide strong evidence that greening is driven by warming (Elmendorf et al., 2012). However, the magnitude of the recent browning is large and cannot be ignored. For both the Eurasian Arctic and the Arctic as a whole, Epstein et al. report the 2014 maxNDVI (greenness) to be below the 33-year average. To find lower values than 2014, you have to go back to 1996 for the whole Arctic and to 1993 for the Eurasian Arctic. However, while browning is the overall trend, there is considerable regional variation and the Arctic is not browning everywhere. These findings raise important questions that represent priority challenges, including (1) what is driving the browning? (2) is browning the new trajectory or only a temporary reversal of greening? and (3) what arctic regions and vegetation types are most sensitive so might show the greatest browning in future? Browning can be caused by a reversal of greening drivers. As raised by Epstein et al. (2015), declines in greenness in some regions, especially early season, arise from greater and longer snow cover (Bieniek et al., 2015), and an earlier analysis showed the declining greening of the Eurasian arctic to be linked to reduced summer warmth index (Bhatt et al., 2013). If so, against an ongoing trend of some of the greatest warming on the planet and declining snow cover duration (AMAP, 2011), we might expect greening to resume soon, although a warmer arctic will result in some areas receiving more snow and having longer snow cover duration (AMAP, 2011). Equally, changes in the ‘trend’ drivers of greening may not explain the widespread extent of browning, and other mechanisms will apply. In particular, while warmer growing seasons (i.e. trend climate change) may be the dominant driver for greening, a number of lines of recent evidence show that extreme events and winter warming drive browning. Critically, both extreme events and winter temperatures are also increasing as part of climate change, and in the case of winter temperatures, increasing much more so than summer (AMAP, 2011). So, despite having the greatest rates of warming globally, the concurrent increases in browning drivers that come with climate change means that we can be far less certain about ongoing greening (example browning and greening scenarios are shown in Fig. 1). For instance, tundra fire can cause complete loss of vegetative cover, and tundra fire frequency is expected to increase with climate change (Bret-Harte et al., 2013). Thermokarst development with permafrost degradation can lead to browning where thaw features expose ground or create water bodies (although succession will re-green these). In the record-low productivity for north-west Scandinavia observed in 2012, 14 different types of anomalous weather events were detected that drove browning (Bjerke et al., 2014). While these occurred further south than the study area of Epstein et al., they may well represent an early warning for higher latitudes as they warm. The best studied of these is currently extreme winter warming which leads to mid-winter bud burst and loss of freeze tolerance (Bokhorst et al., 2011), although the numerous other drivers identified range from warm autumns reducing winter hardening, rain-on-snow resulting in plant ice encasement, lack of snow cover combined with high irradiance leading to frost drought (Fig. 2), through to snow falling on unfrozen ground enhancing snow mould growth and respiratory losses from subnivean plants. In addition, outbreaks of defoliating insects and rust fungi were also cited as browning drivers. Furthermore, events can interact to drive greater browning, such as frost drought damaged heathland becoming more fire prone (Fig. 2f). However, while this may suggest the Arctic faces an onslaught of browning, vegetation also shows evidence of rapid recovery from these events (vegetation in the extreme winter warming and fire studies recovered in 2–4 years; Bokhorst et al., 2011; Bret-Harte et al., 2013). And here lays one of the main challenges. Events are hard to study (and those that occur in winter even more so). The sporadic nature of events in time and space means they cannot be predicted beforehand, infrastructure cannot be set up in advance in the right place and time to monitor an upcoming event, and when they do occur, it may be the aftermath of damage that allows us to detect the event, hence missing the opportunity to study it in progress. The damage is also often transient, with ecosystems recovering after a few years. So, while greening may be viewed more as a steady and gradual process, browning may come from a large number of different drivers, either as trend change but also often as biotic or weather events, reducing greenness in different parts of the landscape, at different times, temporarily. Net browning will arise if the sum of these in space and time overrides greening. Predicting that from a large number of events that we still poorly understand is a real challenge. GKP and JWB are supported by the Leverhulme Trust (UK) (grant F/00 118/AV), the Research Council of Norway (project no. 216434 and 225006) and the Natural Environment Research Council (UK) (grant NE/L002450/1).