Mutual interaction of water column and sediment processes is either neglected or only crudely approximated in many biogeochemical models. We have reviewed the approaches to couple benthic and pelagic biogeochemical models. It is concluded that they can be classified into a hierarchical set consisting of five levels, differing in the amount of detail given to the sediment processes. The most complex approach (level 4) fully couples water column processes to a vertically resolved biogeochemical sediment model. First simplification is achieved by using a vertically integrated dynamic sediment model (level 3); next is a reflective type of boundary (level 2) where particulate material arriving at the sediment surface is instantaneously transformed into dissolved components. Then comes a set of models in which either the bottom-water concentration of dissolved substances or the sediment-water exchange is imposed (level 1). Finally, in some biogeochemical models, the bottom is plainly ignored (level 0). We have tested these various approaches in a coupled physical-pelagic-benthic biogeochemical model for oxygen, nitrogen and carbon cycling in continental shelf areas. We discuss the various model approaches with respect to their impact on the pelagic system and point out some of the inconsistencies hidden in certain formulations. We conclude that lower boundary types, in which sediment fluxes or concentrations are imposed (level 1), are especially badly designed because they fail to assure conservation of mass. Finally, we suggest as best choice a level 3 approach in which the evolution of sedimentary particulate matter is part of the solution and when the bottom fluxes of dissolved constituents are parameterised based on mass budget considerations. These simplified formulations represent the best balance between computational demand and attained accuracy. [KEYWORDS: numerical model; biogeochemical model; early diagenesis; nitrogen cycle; oxygen cycle General-circulation model; organic-matter mineralization; early diagenetic processes; physical-biological model; deep-sea; north-sea; simulation analysis; calcite lysocline; caco3dissolution; silica diagenesis] ; Mutual interaction of water column and sediment processes is either neglected or only crudely approximated in many biogeochemical models. We have reviewed the approaches to couple benthic and pelagic biogeochemical models. It is concluded that they can be classified into a hierarchical set consisting of five levels, differing in the amount of detail given to the sediment processes. The most complex approach (level 4) fully couples water column processes to a vertically resolved biogeochemical sediment model. First simplification is achieved by using a vertically integrated dynamic sediment model (level 3); next is a reflective type of boundary (level 2) where particulate material arriving at the sediment surface is instantaneously transformed into dissolved components. Then comes a set of models in which either the bottom-water concentration of dissolved substances or the sediment-water exchange is imposed (level 1). Finally, in some biogeochemical models, the bottom is plainly ignored (level 0). We have tested these various approaches in a coupled physical-pelagic-benthic biogeochemical model for oxygen, nitrogen and carbon cycling in continental shelf areas. We discuss the various model approaches with respect to their impact on the pelagic system and point out some of the inconsistencies hidden in certain formulations. We conclude that lower boundary types, in which sediment fluxes or concentrations are imposed (level 1), are especially badly designed because they fail to assure conservation of mass. Finally, we suggest as best choice a level 3 approach in which the evolution of sedimentary particulate matter is part of the solution and when the bottom fluxes of dissolved constituents are parameterised based on mass budget considerations. These simplified formulations represent the best balance between computational demand and attained accuracy. [KEYWORDS: numerical model; biogeochemical model; early diagenesis; nitrogen cycle; oxygen cycle General-circulation model; organic-matter mineralization; early diagenetic processes; physical-biological model; deep-sea; north-sea; simulation analysis; calcite lysocline; caco3dissolution; silica diagenesis]