A high resolution spectral wave model is used to quantify the spatial wave climate on geographical scales relevant to intra-site variability for marine renewable energy installations. For the first time, results are compared to in-situ data from an array of four floating wave buoys, and demonstrate the ability of the spectral wave model SWAN (Simulating WAves Nearshore) to resolve spatial differences in the wave climate. Examination of the model source terms highlights bottom friction and refraction as the primary processes contributing to the observed differences across the site. Wave models for climate assessments for marine renewable energy are not commonly operated at sufficient spatial resolution to accurately resolve intra-site variability. This study demonstrates that high spatial resolution spectral wave models, nested into a larger model domain, have the potential to provide an accurate and detailed prediction of the spatial variability of wave conditions across a marine renewable energy site. As such, they could be implemented to provide a more accurate resource assessment for wave energy array deployments, but also for engineering assessments of other marine energy technologies. •Unique validation of a high resolution spectral wave model at marine energy site scale.•Short-term differences in Hm0 regularly exceed 10%, and reach 1.3 m during large sea states.•Differences increase at low frequencies, reaching a long term average of 10% for f = 0.05 Hz–0.1 Hz.•SWAN (Simulating WAves Nearshore) wave model accurately predicts spatial differences in a wave field across a wave energy site.•Application of site-scale spectral model improves accuracy of wave data across a site.