MEPS 251:49-58 (2003)  -  doi:10.3354/meps251049

Geostatistical characterisation of centimetre-scale spatial structure of in vivo fluorescence

Raechel L. Waters*, James G. Mitchell, Justin Seymour

School of Biology, Flinders University, GPO Box 2001, South Australia 5001, Australia

ABSTRACT: We present the first quantitative evidence of 2-dimensional non-random spatial structure for in vivo fluorescence distributions measured at scales of centimetres. To determine the presence of non-random structure and to quantify levels of spatial complexity, 2 common geostatistical analysis techniques, namely Moran¹s I spatial autocorrelation statistic and single-dimension fractal analysis, were employed. Samples were collected at 4 and 2 cm resolution from 2 coastal environments in South Australia; 50% of the distributions were spatially autocorrelated and 60% exhibited fractal scale dependence (with a fractal dimension, DF, ranging from 1.54 to 1.89). Comparison of distributions sampled in the horizontal and vertical indicate that structural complexity is equally important in both planes. Structural complexity was greater at 2 cm than at 4 cm, suggesting that structure may arise at or below the 2 cm scale. This observation is in direct conflict with the general assumption that turbulence homogenises distributions at these scales.


KEY WORDS: Two-dimensional sampling · Moran¹s I-statistic · Correlogram · Fractal analysis


Full text in pdf format