Terrestrial　laser　scanning　(TLS)　is　a　promising　tool　in　geodetic　applications.　Data　quality　of　outdoor　TLS　measurements　is　considerably　limited　by　external　errors　caused　by　atmospheric　variations　and　object　related　factors.　In　this　study,　external　models　related　to　atmospheric　refraction,　beam　wander　and　incidence　angle　were　introduced　and　integrated　into　a　seven-parameter　similarity　transformation　(7PST)　model　to　compose　a　combined　model.　The　combined　model　was　rigorously　adjusted　by　an　iterative　Gauss-Helmert　(GH)　model　primarily　to　estimate　external　calibration　parameters　and　transformation　parameters　simultaneously.　In　the　application　on　a　dam　surface,　a　t-test　showed　that　a　significant　external　calibration　parameter　was　the　range　scale,　which　mainly　corresponded　with　the　incidence　angle　model.　The　precision　of　estimated　representative　points　was　improved　by　approximately　50%　with　the　proposed　method.　The　presented　models　show　a　great　potential　for　detecting　external　errors　and　registering　multiple　scans　in　outdoor　applications.