Over 50 million people in South Asia are exposed to groundwater contaminated with carcinogenic arsenic(III). Photocatalyst-adsorbent composite materials are popularly developed for removing arsenic in a single-step water treatment. Here, As(III) is oxidised to As(V), which is subsequently removed via adsorption. We previously developed a component additive surface complexation model (CA-SCM) to predict the speciation of arsenic adsorbed onto TiO2/Fe2O3 under different environmental conditions, using surface complexes taken from studies of single-phase minerals. In this work, we critically evaluate this approach, using experimental observations of the surface structures of arsenic adsorbed onto TiO2/Fe2O3. Extended X-ray absorption fine structure spectroscopy (EXAFS) indicates significant As(III) surface precipitation, and the possible formation of tridentate 3C complexes. EXAFS was unable to identify As binding modes for TiO2 and Fe2O3 surface complexes simultaneously, highlighting the challenge of analysing composite surfaces. FTIR and zeta potential analysis indicate that As(III)-Fe2O3 surface complexes are protonated at neutral pH, whilst As(III)-TiO2, As(V)-Fe2O3 and As(V)-TiO2 surface complexes are negatively charged. Our study confirms the speciation predicted by CA-SCM, particularly As(III) surface precipitation, but also introduces the possibility of tridentate As(III) at acidic pH. This study highlights how experiment and modelling can be combined to assess surface complexation on composite surfaces.