The species richness-productivity relationship varies among regions and productivity estimates, but not with spatial resolution

The relationship between species richness and productivity (SRPR) has been
a long-studied and hotly debated topic in ecology. Different studies have
reported different results with variable shapes (i.e. unimodal, linear)
and directions (i.e. positive, negative) of SRPRs depending on spatial
grain (i.e. size of sampling unit for species richness), productivity
estimates, and study extent. In this study, we quantified the effect of
multiple estimates of productivity (aboveground, belowground and total
biomass, and various measures of soil fertility) on species richness
across three spatial grains (0.04 m2, 1 m2, and 25 m2) across temperate
grasslands from two regions in Central Europe. We analyzed SRPR in each of
the two regional datasets separately, as well as the two datasets pooled
together. Our results have revealed that differences caused by spatial
grain were unexpectedly small, and the direction of the SRPR was
consistent within each productivity estimate, but differed between
regions. Productivity estimates (across all spatial scales) had different,
sometimes contrasting effects on SRPR (together with predictive power)
within a region, and this pattern was more pronounced when compared
between regions. The combination of different datasets led to very
different results than when these were analyzed separately. We did not
find any evidence for a unimodal response. This study points to the
necessity of careful assessing when combining datasets from different
regions, even if the plant communities belong to the same vegetation type.
The dataset combination may blur the role of different drivers, which
likely determine the shape and strength of SRPR. We suggest that data and
study comparability may be enhanced by consistently using the same
productivity estimates, which would allow for more robust interpretation
of possible ecological drivers underlying the SRPR.