Affected　by　discontinuities,　the　properties　of　rock　masses　are　characterized　by　strong　scale　dependency　and　anisotropy.　Rock　mass　samples　taken　at　any　scale　smaller　than　the　representative　elementary　volume　(REV)　size　could　lead　to　an　incorrect　characterization　and　property　upscaling.　To　better　understand　the　sampling　problem,　numerical　tests　based　on　an　outcrop-data-based　discrete　fracture　network　(DFN)　model　were　conducted,　trying　to　determine　the　REV　size　and　its　anisotropy.　The　model　was　validated　and　subsequently　sampled　to　produce　455　rectangular　samples　with　a　width　ranging　from　0.05　to　21　m　and　a　constant　height-to-width　ratio　of　2.　The　samples　were　introduced　into　a　3D　particle　flow　code　model　to　create　synthetic　rock　mass　(SRM)　samples.　Numerical　uniaxial　compressive　tests　at　different　loading　directions　were　performed　to　study　the　scale　dependency　and　anisotropy　of　mechanical　parameters.　The　results　show　that　the　mechanical　REV　sizes　in　different　directions　differ,　changing　between　7　and　19　m.　The　mechanical　properties　of　rock　mass　samples　in　a　REV　size　exhibit　strong　anisotropy,　with　the　values　of　uniaxial　compressive　strength　(UCS)　and　elastic　modulus　(E)　varying　from　5.6　to　10.3　MPa　and　3.9　to　8.0　GPa,　respectively.　The　simulated　values　of　UCS　were　validated　based　on　GSI　and　the　Hoek-Brown　failure　criterion.　The　geometrical　REV　size　based　on　the　volumetric　fracture　intensity　was　calculated　to　be　7　m,　equal　to　the　minimum　of　the　mechanical　REV　size;　this　suggests　that　the　geometrical　REV　defines　the　lower　bound　of　the　REV　size.