Jeison Sosa, Christopher Sampson, Andrew Smith, Jeffrey Neal, Paul Bates

Hydrodynamic floodplain inundation models have been popular for many years and used extensively in engineering applications. Continental scale flood studies are now achievable using such models due to the development of terrain elevation, hydrography and river width datasets with global coverage. However, deploying flood models at any scale is time-consuming since input data needs to be processed from different sources. Here we present LFPtools, which is an open-source Python package which encompasses most commonly used methods to prepare input data for large scale flood inundation studies using the LISFLOOD-FP hydrodynamic model. LFPtools performance was verified over the Severn basin in the UK where a 1 km flood inundation model was built within 1.45 min. Outputs of the test case were compared with the official flood extent footprint of a real event and satisfactory model performance was obtained: Hit rate = 0.79, False alarm ratio = 0.24 and Critical success index = 0.63.