S. Lovejoy et al., Universal multifractals and ocean patchiness: phytoplankton, physical fields and coastal heterogeneity, J PLANK RES, 23(2), 2001, pp. 117-141
We argue that a wide-range scaling approach is demanded by standard Stommel
diagrams and that it can unify the treatment of phytoplankton variability
over wide ranges of scales. By investigating the effects of coastal heterog
eneity on the variability of in situ salinity (S), oxygen (rho (O)), temper
ature (T), optical transmissivity (t) and phytoplankton proxy data (fluores
cence; rho (p)) over the range similar to0.4-1600 m, we statistically chara
cterize the heterogeneity of these variables, determining both the range an
d types of scaling, as well as their scale-by-scale interrelationships. By
comparing bays with systematically varying large-scale heterogeneity, we we
re able to investigate the influence of the latter on the variability, syst
ematically determining the three universal multifractal parameters as well
as the exponent characterizing extreme self-organized critical behaviour. W
e found that, consistent with turbulent dominated dynamics, T, rho (O), S a
nd t were scaling over essentially the entire observed range of scales, wit
h T and rho (O) being statistically very close to passive scalars. However,
rho (p) was quite different, displaying two regimes separated by a charact
eristic 'planktoscale' typically similar to 100 m, but highly variable. The
large-scale regime was neither passive scalar nor growth dominated (Denman
-Platt), but was rather in between the two (the corresponding exponent was
H-p approximate to 1/8 rather than 0 or 1/3, respectively). In addition, we
found a new small-scale regime with H-p approximate to -1/3, which is much
'rougher' than passive scalar (which has H-p = +1 /3). We propose a simple
model involving both growth and turbulence to account for the large scale,
and grazing and turbulence (predator-prey zooplankton/phytoplankton intera
ctions) to account for the small scale. Depending on the value of a dimensi
onless grazing constant Gr = D/(tau (2)(g)epsilon) (where D is the zooplank
ton diffusion constant, tau (g) is the phytoplankton growth constant and ep
silon is the turbulent energy flux), the small scale is dominated either by
the turbulent grazing (Gr >1) or by passive scalar turbulence (Gr < 1). In
the grazing regime, we also theoretically predict that the density fluctua
tion exponent = -1/3, which is quite close to the data and quantifies the t
endency of the zooplankton to uniformize the phytoplankton distribution by
preferentially grazing high-concentration patches.