Du. Hooper et L. Johnson, Nitrogen limitation in dryland ecosystems: Responses to geographical and temporal variation in precipitation, BIOGEOCHEMI, 46(1), 1999, pp. 247-293
We investigated the relationship between plant nitrogen limitation and wate
r availability in dryland ecosystems. We tested the hypothesis that at lowe
r levels of annual precipitation, aboveground net primary productivity (ANP
P) is limited primarily by water whereas at higher levels of precipitation,
it is limited primarily by nitrogen. Using a literature survey of fertiliz
ation experiments in arid, semi-arid, and subhumid ecosystems, we investiga
ted the response of ANPP to nitrogen addition as a function of variation in
precipitation across geographic gradients, as well as across year-to-year
variation in precipitation within sites. We used four different indices to
assess the degree of N limitation: (1) Absolute Increase of plant productio
n in response to fertilization (the slope of ANPP vs. amount of added N at
different levels of annual precipitation); (2) Relative Response (the perce
nt increase in fertilized over control ANPP at different levels of N additi
on); (3) Fertilizer Use Efficiency (FUE, the absolute gain in productivity
per amount of fertilizer N), and (4) Maximum Response (the greatest absolut
e increase in ANPP at saturating levels of N addition). Relative Response t
o fertilization did not significantly increase with increasing precipitatio
n either across the geographic gradient or across year-to-year variation wi
thin sites. Nor did the Maximum Response to fertilization increase with inc
reasing precipitation across the geographic gradient. On the other hand, th
ere was a significant increase in the Absolute Increase and FUE indices wit
h both geographical and temporal variation in precipitation. Together, thes
e results indicate that there is not necessarily a shift of primary limitat
ion from water to N across the geographic water availability gradient. Inst
ead, our results support the hypothesis of co-limitation. The apparently co
ntradictory results from the four indices of N limitation can best be expla
ined by an integration of plant ecophysiological, community, and ecosystem
mechanisms whereby plants are co-limited by multiple resources, species shi
fts occur in response to changing resource levels, and nitrogen and water a
vailability are tightly linked through biogeochemical feedbacks.