Geographical variability of environmental parameters versus GPS precision: toward a better sampling strategy

To characterize a sedimentary environment, it is risky to take a single sample when the spatial variability is unknown. A reference station has to reflect the natural variations in order to allow the creation of long time series. However, it can remain unclear whether the temporal changes are real or due to a spatial variation. We highlight here the importance of spatial variability at the scale of precision of the GNSS. It appears that the number and arrangement of replicates depend on the environment and the studied parameters. InC, TOC and TS show a sufficiently low spatial variability to allow temporal tracking using GNSS without multiplying samples. The fine fraction percent shows a high spatial variability over small distances. The study of this parameter in the framework of temporal tracking requires a knowledge of its spatial variability during each period of sampling, and hence leads to the multiplication of samples.