We present and interpret Global Positioning System (GPS) measurements of crustal motions for the period 1988–1997 at 189 sites extending east‐west from the Caucasus mountains to the Adriatic Sea and north‐south from the southern edge of the Eurasian plate to the northern edge of the African plate. Sites on the northern Arabian platform move 18±2 mm/yr at N25°±5°W relative to Eurasia, less than the NUVEL‐1A circuit closure rate (25±1 mm/yr at N21°±7°W). Preliminary motion estimates (1994–1997) for stations located in Egypt on the northeastern part of Africa show northward motion at 5–6±2 mm/yr, also slower than NUVEL‐IA estimates (10±1 mm/yr at N2°±4°E). Eastern Turkey is characterized by distributed deformation, while central Turkey is characterized by coherent plate motion (internal deformation of <2 mm/yr) involving westward displacement and counterclockwise rotation of the Anatolian plate. The Anatolian plate is de‐coupled from Eurasia along the right‐lateral, strike‐slip North Anatolian fault (NAF). We derive a best fitting Euler vector for Anatolia‐Eurasia motion of 30.7°± 0.8°N, 32.6°± 0.4°E, 1.2°±0.1°/Myr. The Euler vector gives an upper bound for NAF slip rate of 24±1 mm/yr. We determine a preliminary GPS Arabia‐Anatolia Euler vector of 32.9°±1.2°N, 40.3°±1.1°E, 0.8°±0.2°/Myr and an upper bound on left‐lateral slip on the East Anatolian fault (EAF) of 9±1 mm/yr. The central and southern Aegean is characterized by coherent motion (internal deformation of <2 mm/yr) toward the SW at 30±1 mm/yr relative to Eurasia. Stations in the SE Aegean deviate significantly from the overall motion of the southern Aegean, showing increasing velocities toward the trench and reaching 10±1 mm/yr relative to the southern Aegean as a whole.

SUMMARY We use continuously recording GPS (CGPS) and survey‐mode GPS (SGPS) observations to determine Euler vectors for relative motion of the African (Nubian), Arabian and Eurasian plates. We present a well‐constrained Eurasia–Nubia Euler vector derived from 23 IGS sites in Europe and four CGPS and three SGPS sites on the Nubian Plate (−0.95 ± 4.8°N, −21.8 ± 4.3°E, 0.06 ± 0.005° Myr−1). We see no significant (>1 mm yr−1) internal deformation of the Nubian Plate. The GPS Nubian–Eurasian Euler vector differs significantly from NUVEL‐1A (21.0 ± 4.2°N, −20.6 ± 0.6°E, 0.12 ± 0.015° Myr−1), implying more westward motion of Africa relative to Eurasia and slower convergence in the eastern Mediterranean. The Arabia–Eurasia and Arabia–Nubia GPS Euler vectors are less well determined, based on only one CGPS and three SGPS sites on the Arabian Plate. The preliminary Arabia–Eurasia and Arabia–Nubia Euler vectors are 27.4 ± 1.0°N, 18.4 ± 2.5°E, 0.40 ± 0.04° Myr−1, and 30.5 ± 1.0°N, 25.7 ± 2.3°E, 0.37 ± 0.04° Myr−1, respectively. The GPS Arabia–Nubia Euler vector differs significantly from NUVEL‐1A (24.1 ± 1.7°N, 24.0 ± 3.5°E, 0.40 ± 0.05° Myr−1), but is statistically consistent at the 95 per cent confidence level with the revised Euler vector reported by Chu & Gordon based on a re‐evaluation of magnetic anomalies in the Red Sea (31.5 ± 1.2°N, 23.0 ± 2.7°E, 0.40 ± 0.05° Myr−1). The motion implied in the Gulf of Aqaba and on the Dead Sea fault (DSF) by the new GPS Nubia–Arabia Euler vector (i.e. ignoring possible Sinai block motion and possible internal plate deformation) grades from pure left lateral strike‐slip in the Gulf and on the southern DSF with increasing compression on the central and northern DSF with relative motion increasing from 5.6 to 7.5 mm yr−1 (±1 mm yr−1) from south to north. Along the northern DSF (i.e. north of the Lebanon restraining bend) motion is partitioned between 6 ± 1 mm yr−1 left‐lateral motion parallel to the fault trace and 4 ± 1 mm yr−1 fault‐normal compression. Relative motions on other plate boundaries (including the Anatolian and Aegean microplates) derived from the GPS Euler vectors agree qualitatively with the sense of motion indicated by focal mechanisms for large crustal earthquakes (M > 6). Where data are available on fault‐slip rates on plate bounding faults (North Anatolian fault, East Anatolian fault, Dead Sea fault, Red Sea rift), they are generally lower than, but not significantly different from, the full plate motion estimates suggesting that the majority of relative plate motion is accommodated on these structures.

We present and interpret Global Positioning System (GPS) measurements of crustal motions for the period 1988–1997 at 189 sites extending east‐west from the Caucasus mountains to the Adriatic Sea and north‐south from the southern edge of the Eurasian plate to the northern edge of the African plate. Sites on the northern Arabian platform move 18±2 mm/yr at N25°±5°W relative to Eurasia, less than the NUVEL‐1A circuit closure rate (25±1 mm/yr at N21°±7°W). Preliminary motion estimates (1994–1997) for stations located in Egypt on the northeastern part of Africa show northward motion at 5–6±2 mm/yr, also slower than NUVEL‐IA estimates (10±1 mm/yr at N2°±4°E). Eastern Turkey is characterized by distributed deformation, while central Turkey is characterized by coherent plate motion (internal deformation of <2 mm/yr) involving westward displacement and counterclockwise rotation of the Anatolian plate. The Anatolian plate is de‐coupled from Eurasia along the right‐lateral, strike‐slip North Anatolian fault (NAF). We derive a best fitting Euler vector for Anatolia‐Eurasia motion of 30.7°± 0.8°N, 32.6°± 0.4°E, 1.2°±0.1°/Myr. The Euler vector gives an upper bound for NAF slip rate of 24±1 mm/yr. We determine a preliminary GPS Arabia‐Anatolia Euler vector of 32.9°±1.2°N, 40.3°±1.1°E, 0.8°±0.2°/Myr and an upper bound on left‐lateral slip on the East Anatolian fault (EAF) of 9±1 mm/yr. The central and southern Aegean is characterized by coherent motion (internal deformation of <2 mm/yr) toward the SW at 30±1 mm/yr relative to Eurasia. Stations in the SE Aegean deviate significantly from the overall motion of the southern Aegean, showing increasing velocities toward the trench and reaching 10±1 mm/yr relative to the southern Aegean as a whole.

GPS survey sites in the Sinai Peninsula show northerly motion relative to Africa (Nubia) at 1.4 F 0.8 mm/yr north and 0.4 F 0.8 mm/yr west. Continuous IGS GPS sites in Israel, west of the Dead Sea fault show a similar northerly sense of motion relative to Nubia (2.4 F 0.6 mm/yr north and 0.04 F 0.7 mm/yr east), suggesting that the entire Sinai Block south of Lebanon is characterized by northward translation relative to the Nubian plate. We develop an elastic block model constrained by the GPS results that is consistent with the regional tectonics and allows us to estimate slip rates for Sinai bounding faults, including the Gulf of Aqaba-southern Dead Sea fault system (~4.4 F 0.3 mm/yr, left lateral), the Gulf of Suez (1.9 F 0.3 mm/yr left lateral, and 1.5 F 0.4 mm/yr extension), and the Cyprus Arc (predominantly convergence at 8.9 F 0.4 mm/yr along the western segment, and~6.0 F 0.4 mm/yr left lateral, strike slip along the eastern segment). These observations imply that the Sinai Peninsula and Levant region comprise a separate sub-plate sandwiched between the Arabian and Nubian plates.