Attempts to increase European biodiversity by restoring rivers and floodplains are based on inadequate data on natural systems. This is particularly the case for NW European rivers because all catchments have been impacted by agriculture and river engineering. If river restoration is to have an ecological, as opposed to `cosmetic' design basis then baseline models are required. However, this poses three questions; (a) what is the natural river-floodplain state, (b) how can it be defined and modelled and (c) can this state be recreated today? The first two questions can only be addressed by using palaeohydrological and palaeoecological data. A second and equally vital consideration is the stability/instability of any restored system to change in external forcing factors (e.g. climate) and in this context it may not be realistic to expect baseline models to provide equilibrium solutions but instead to define process-form domains. Over the last two decades evidence has accumulated that the natural state of lowland rivers in much of NW Europe was multi rather than single thread-braided, anastomosing or anabranching. Until recently our knowledge of floodplain palaeoecology was generally derived from pollen diagrams, which have source-area of problems and lack of taxonomic specificity. The precision and breadth of palaeoecological reconstruction (including richness and structure) has been greatly increased by the use of multiple palaeo-indicators including macrofossils, diatoms and beetles. The dynamics of small to medium sized, low-energy, predeforestation floodplains were dominated by disturbance (windthrow, beavers, etc.) and large woody debris. In order to compare the hydrogeomorphological basis of floodplain ecology, both temporally and spatially, a simple index of fluvial complexity is presented. Palaeoecological and geomorphological investigations have the potential to provide in-depth models of the natural range of channel conditions and sensitivity to external change that can be used to provide a scientific basis for floodplain restoration. There is also the possibility that floodplain-channel restoration may be a valuable tool in the mitigation of future geomorphological change forced by climatic instability.