Loss on ignition as a method for estimating organic and carbonate content in sediments: reproducibility and comparability of results