Distribution of suspended particulate matter in the North Sea as inferred from NOAA/AVHRR reflectance images and in situ observations
Van Raaphorst, W.; Philippart, C.J.M.; Smit, J.P.C.; Dijkstra, F.J.; Malschaert, J.F.P. (1998). Distribution of suspended particulate matter in the North Sea as inferred from NOAA/AVHRR reflectance images and in situ observations. J. Sea Res. 39(3-4): 197-215. https://dx.doi.org/10.1016/s1385-1101(98)00006-9
In: Journal of Sea Research. Elsevier/Netherlands Institute for Sea Research: Amsterdam; Den Burg. ISSN 1385-1101; e-ISSN 1873-1414, more
|  |
| Keyword |
|
| Author keywords |
North Sea; suspended matter; remote sensing; in situ data; interpolation |
| Authors | | Top |
- Van Raaphorst, W.
- Philippart, C.J.M., more
- Smit, J.P.C.
|
- Dijkstra, F.J.
- Malschaert, J.F.P.
|
|
| Abstract |
Distribution patterns of suspended particulate matter (SPM) in the surface water of the North Sea were calculated on the basis of: (1) the 1973–1993 data base of the EC MAST North West European Shelf Programme (NOWESP); and (2) composite reflectance images constructed from data that were collected by the NOAA/AVHRR satellite in 1990–1991. Three models were used for interpolating the in situ data: (1) a distance-weighted interpolation algorithm in which only the in situ data are taken into account; (2) an algorithm in which the ratios between the measured SPM concentrations and reflectances are interpolated, and the distribution of SPM is calculated from the field of interpolated ratios and the synoptic reflectance image; and (3) a distance-weighted algorithm similar to model-1, but with an additional weight factor that is based on local differences in reflectance. The models were tested for periods of 1 and 3 weeks in September 1990 and January 1991, and for the merged set consisting of all in situ data measured in September and January, respectively, between 1973 and 1993. Model-2 and -3 gave largely similar results and had a performance superior to model-1, particularly because they showed more detailed structures in the spatial distributions. Validations and cross-validations showed that the absolute concentrations of SPM predicted by the models were too low at high in situ concentrations and too high at low in situ concentrations. This shortcoming was due to the relatively high degree of smoothing that we applied in the models to account for the large variance of the in situ data. Semivariograms and correlograms indicated that the in situ data had substantial variability and were poorly correlated even at short distances. Only for the 20-year-merged data set did some correlation (<60%) exist for stations <50 km from each other. Monthly distributions of SPM were calculated with model-3 and the 20-year data set. The distributions confirm the main patterns previously found by others, such as the turbidity plume crossing the North Sea from southeast England towards the depository in the Skagerrak and the Norwegian Channel. The distributions indicate that materials from this plume may be deposited in the central North Sea in spring and summer and eroded again in autumn and winter. Areas with maximum SPM concentrations were identified off the Belgian coast and north of the Wadden Sea, particularly in winter, from which particles are entrained into the main current in a narrow strip along the continental coast to the German Bight. The results suggest that the two main fluxes of SPM in the North Sea, off England and along the continental coast, remain largely separated until they both end in the Skagerrak. |
|
