DEPARTURE FROM QUARTER-POWER SCALING OF AQUATIC HABITAT RICHNESS-AREA AS A STRUCTURAL INDICATOR OF CONSTRAINTS ON RIVERSCAPE ORGANIZATION

The complexity of ecological systems arises from interactions among a variety of different components at different hierarchical levels (for example: genes to landscapes). Yet despite such complexity, most systems exhibit regularities that take the form of scaling laws. Such laws can provide a measure of ecosystem organization and give insights into the underlying mechanisms that structure ecosystems (anthropogenic or other).

In this study, we tested the hypothesis that aquatic habitat richness of the Upper Mississippi (UMR) and Illinois rivers (IR) can be quantified by the power function R=cA^z which predicts an increase in aquatic habitat richness (R) with increasing area (A). We also hypothesized that variation in c and z would be associated with modifications to the geometry of the river and floodplain (i.e. river engineering). Using photo-interpreted maps of the distribution of 11 aquatic habitats, we centered focal landscapes (i.e. windows) of various sizes (1, 2, 5, 10, 15, 25, 50, and 100 ha) on each of 19.3 million 10m aquatic pixels, estimated habitat richness within each landscape, and summarized the results for four floodplain reaches and 107 management units of the UMR and IR.

R² values exceeded 0.93 for all four floodplain reaches and 107 management units of the river system and were >0.97 for all reaches and 96 units. Estimates of c ranged from 1.0 in areas with a single habitat type within 1 ha landscapes to 2.5 in areas with more than two habitat types within 1 ha landscapes. Estimates of z ranged from 0.11 where habitat richness increased slowly with increasing landscape size to ~0.25 where habitat richness increased more quickly with landscape size. Variation in c and z was associated with modifications to the geometry of the rivers and floodplains. Estimates of c were homogeneously low (<1.6) where % of the floodplain impounded by locks-and-dams exceeded 10%. Estimates of z were lowest (~0.11) where impoundment exceeded 60% of the floodplain and increased to ~0.25 with increases in the % area in secondary channels. Thus, variation in the parameter estimates c and z appear to be useful coarse-scale indicators of the impacts of river engineering on the structural diversity of the Upper Mississippi and Illinois Rivers. We propose that departure of the scaling parameter (z) from 0.25 may reveal constraints on land and riverscape organization more generally.