The newly named DendroGeeks working group at the I&P workshop has now completed our data analysis! We sampled 2 Ulmus americana (American Elm) trees in each of three watersheds at Konza Prairie. The watersheds are all bison-grazed, but experience different burn regimes: every two years (N2B), every four years (N4C), and annual burns in the 90s, switching to every 20 years in 1993 (N20A). See this map for details of grazing and burn regimes at Konza.
We measured ring widths on all trees from 1990-2010, and ran nitrogen isotope analysis on years 2002-2006 (for N2B and N4C trees), and 1992-1993, 2002-2004 (for N20A).
As you can see in Figure 1, raw ring width decreased similarly with age for all trees in all watersheds, which is expected under normal growing conditions.
We then compared ring width and nitrogen isotopic signature (abbreviated δ15N) in burn years, one year immediately post burn, and any years beyond one year post burn (Figs 2 and 3). Ring width showed no significant trends, but saw a slight increase in years beyond burns (Fig 2). δ15N was significantly higher in burn years, and all groups showed depleted (less than 0) δ15N values (Fig 3). Higher δ15N in tree rings may reflect increased ammonia volatilization (the loss of nitrogen from the ecosystem in the form of NH3 gas) due to increased ground temperatures during burning, which results in higher δ15N of the remaining nitrogen pool, increasing tree ring δ15N values.
Perhaps most interesting was the comparison between ring width and δ15N between burn, post, and beyond years (Figure 4). In this comparison, burn and post-burn years exhibit no relationship between ring width and δ15N, while rings more than two years beyond burn events showed a strong negative trend between ring width and δ15N. The group then looked more closely at the beyond-burn years, which included years from both 4-year and 20-year burn treatments, spanning two to ten years post burn event. Interestingly, 2-3 year post burn years (from watershed N4C) showed a slight increasing trend and were significantly higher than 10-12 year post burn years (from watershed N20A) which showed a slight decreasing trend. Unfortunately, our sampling strategy confounds the interpretation of these data: it could be that short-term post burn relationships differ significantly from longer-term post burn trends. It could also be that watershed N4C exhibits fundamentally different nitrogen cycling than watershed N20A. In order to tease apart these two issues, the group could (if time and budgets allowed…) sample rings 2-3 years post burn in the N20A watershed, and sample additional trees from other 4-year and 20-year burn regime watersheds at Konza.
These data show interesting trends between tree growth and nitrogen cycling between different disturbance regimes, and could form a solid basis for a longer-term and more intensive analysis of nitrogen cycling at Konza Prairie watersheds. Congrats DendroGeeks on a job well done!!