Arctic LTER

20-year long Nitrogen Addition Experiment Causes Functional Changes to Arctic Ecosystems

Fig. 1. Aboveground net primary production in fertilized and control plots in moist acidic tussock tundra at Toolik Lake, 1982-2000

Industrialization increases the deposition of atmospheric nitrogen worldwide. In the Arctic, the amount eventually will double, although deposition will be less in our study areas than in areas directly downwind from industrial areas. Repeated observation and sampling of long-running experiments allows us to determine whether the initial magnitude and trajectory of ecosystem response to this kind of environmental change is sustained over the long term, and to examine interactions among slow- and fast-responding components of overall ecosystem change. Repeated observation also allows us to interpret long-term change in response to our manipulations in the context of long-term "normal" or "background" variation in ecosystem states and processes. For example, in the summer of 2000 we completed our fifth harvest of a 20-year old fertilizer experiment (the 6th harvest of control plots for this experiment). After 20 years it was clear that fertilized plots were distinctly different from control plots, not only in terms of their total production and biomass, but that they also differed functionally in terms of the relationships between productivity, leaf mass, and leaf area (Fig. 1). This is significant because as Arctic soils warm and the permafrost thaws deeper, there will be more nutrients available to the plants from the soil.

Remarkably, leaf mass in the fertilized plots was significantly lower than in the controls in both 1995 and 2000, while productivity in the fertilized plots was more than double that of controls. How is this possible? The main reason for this result is that the fertilized plots are now strongly dominated by dwarf birch, Betula nana, which has much thinner leaves (i.e., a much higher specific leaf area). Because birch has thin leaves, it can produce more than twice the leaf area of controls with a lower total leaf mass. This leads also to very different overall pattern of canopy and aboveground N allocation, with a much more efficient photosynthetic return per unit canopy N (Shaver et al. 2001). It also allows much greater proportional allocation to woody stem growth, greatly increasing both the height of the canopy and the total aboveground biomass (Bret-Harte et al. 2002). This interpretation of the causes and implications of a change in plant functional type composition is radically different from earlier hypotheses (e.g., Chapin and Shaver 1985) and would not have been possible without long-term experimental evidence.

Mobilization of nutrients resulting from human activities and global climate change will also affect stream ecology

At one time, ARC LTER investigators believed that the low productivity of the Kuparuk River was due to a natural lack of nutrients rather than harsh winter conditions, cold temperatures and physical scouring of the rocky stream bottom. To test this theory, we started in 1983 a long-term experimental fertilization study of the Kuparuk River at an amount that is five-to-10-times greater than the natural levels of phosphate entering the river. We wanted the amount to be high enough so that it would have a measurable effect, but not so high that it would turn the stream bottom green. Even in the first summer of the fertilization it was apparent that phosphorus addition during the summer stimulated biotic activity of all types in the river (Peterson et al.1985). Bacterial activity on both labile and refractory substrates was increased by phosphorus addition (Hullar and Vestal 1989). Diatoms on the rocks responded within a few days to the added phosphorus and within a few years insect abundance and fish growth were measurably affected by the addition (Peterson et al. 1993). Some insects such as black flies declined but others including Baetis and Brachycentrus increased (Hershey and Hiltner 1988, Hershey et al.1988), contributing to accelerated fish growth (Deegan and Peterson 1992). After several more years of fertilization we debated stopping the experiment because we thought we might have already observed the full response. However we decided to continue the experiment and to our surprise moss began to proliferate on rocks in riffles after 8 or 9 years of continuous summer fertilization (Bowden et al. 1992). Over the next several years a dense mat of the moss Hygrohypnum covered the majority of what had previously been bare rock habitat covered by a thin biofilm of bacteria and diatoms. Again the insect populations underwent dramatic changes with increases in chironomids and Ephemerella and declines in Orthocladius and Baetis. The effect of this recent transition in community structure on grayling is still unknown although surprisingly it does not yet appear to be large. A second stream fertilization was performed in Oksrukuyik Creek starting in 1991. The pattern of response was similar to that observed in the Kuparuk with the exception that filamentous algae were more abundant in the fertilized reach than was observed in the Kuparuk (Harvey et al. 1998).

Figure. 2 July water temperature at 1 m for each year at Toolik Lake, Arctic LTER

Figure 3. Non-linear alkalinity increase of Toolik Lake water

Figure 4.The reduction in size of lake trout in theToolik Lake study

Long-term Lake Research Illustrates Effects of Climate Change and Human Impact

Far and away the longest monitored lake is Toolik Lake, which was first studied in 1975 (O'Brien et al. 1997). Over time we have noted a variety of long-term changes. One is the warming of the mean July water temperature at a depth of 1 meter, for each year (Fig. 2); the overall trend is statically significant with the regression having a p value of 0.025. The "y" intercepts for 1975 versus 2000 indicate that, on average, the lake is two-degrees C warmer now than at the beginning of the study (Hobbie et al. 1999). While even 25 years is too short a time to confirm Arctic warming, these data are certainly consistent with such a warming trend.

We have also noticed that the alkalinity of Toolik Lake water has increased (Fig. 3). This increase has not been linear but shows two spurts: one early in our study of the lake and then one much more recently with a cubic equation fitting the data with a remarkably low p value. Analysis of the precipitation chemistry shows that this increase is not due to increased salinity in the precipitation but is most likely coming out of the Toolik watershed, possibly due to deepening of the thaw layer above the permafrost.

Another long-term trend in Toolik Lake we have noted is the reduction in size of lake trout. Toolik Lake is primarily an experimental site, but during production of the Alaskan oil pipeline, it did experience an episode of heavy recreational fishing. Early in the study in 1977, the median weight of lake trout was 578 g and by 1986 this was reduced to almost one-half of that (Fig. 4) (McDonald and Hershey 1989). By 1997 the median weight of lake trout was still 60% of the 1977 weight. This reduction in the size is doubtless due to the episode of heavy fishing pressure correlated with the pipeline activity Toolik Lake has received, and to the very slow growth of lake trout. These fish live to be 50-60 years old, and we estimate that the growth rate for lake trout averages a less than five-percent increase per year for adult fish.

Whole-lake Experimentation Provides Window to Global Change

Figure 5. The mean annual chlorophyll a density for the upper three meters of the sectored Lake N-2. The open bars represent the treated sector of the lake while the solid bars represent the reference sector of the lake. The asterisks under each year give an indication of the level of significant difference between the treated and reference sector for that year.

Fig. 6. Sediment flux of inorganic nitrogen and phosphorus from the treated sector of Lake N-2. The open bars represent the flux of nitrogen while the solid bars represent the flux

Experiments that amplify conditions in nature to those that could occur with current levels of human activity provide investigators with insights to global change.

At the Arctic LTER site on the North Slope of Alaska, we have gained a great deal of understanding of the functioning of Arctic lakes through a series of whole-lake manipulations. One of the most successful experiments has been dividing a lake with a polyethylene curtain and adding inorganic nitrogen and phosphorus for six summers to the downstream sector of the lake (Hershey 1992). The nutrient loading levels were 2.91 milimols of nitrogen per square meter per day and 0.23 mmols of phosphorus per square meter per day, which is five times the estimated nutrient loading of Toolik Lake. A five-fold increase in nutrients is a realistic amount that could occur with current levels of human activity.

The phytoplankton, as measured by chlorophyll a, responded positively to the nutrient addition, which began in 1985; each year the treated sector had significantly more phytoplankton than the reference sector (Fig. 5). However, for the first four years of the experiment there was no carry-over from one year to the next. That is, in the early summer both sectors looked the same and indeed had virtually identical chlorophyll levels even though the chlorophyll levels may have differed by as much as eight times at the end of the previous summer. The reason for this was that during the first four summers none of the added phosphorus was recycled from the sediments. This was verified through the use of benthic chambers, which can measure nutrient fluxes from the sediment (Fig. 6). Some nitrogen did recycle early in the experiment but it was not until the fifth year of the experiment that any phosphorus was released from the sediment of the treated sector lake N-2.