Crustal stability is a key factor for selection of engineering construction sites (Hu et al., 1984; Yin et al., 1990, 1992). Statistical analyses and artificial neural networks are used for calculation of the present tectonic activity and the quantitative assessment of crustal stability along the Golmud-Lhasa Railway. Tectonic activity (α≤1)and crustal stability β(≤1) are two important functions in the classification and selection of engineering sites, where α=1−β. The strength of the present tectonic activity (α_i) may be evaluated in the Tibetan Plateau by using four available factors: slip-rate of active fault (V_i), magnitude of earthquake (M_i), temperature of hotspring (T_i) and tectonic strain (ε_i) for any unit (i). Their relation for tectonic activity is α_i = (V_i/V_max + M_i/M_max +T_i/T_max +ε_i/ε_max) /α_max.

The strength of tectonic activity (α_i) can be contoured automatically by using computer and related software. This provides data for a quantitative assessment of tectonic activity and classification of engineering stability. It is shown that unstable tectonically active regions with α≥0.40 may produce Ms≥6-7 earthquakes, and extreme unstable tectonically active region with α≥0.70 may produce Ms≥8 earthquakes (Wu et al., 2005). Seven unstable regions, the South Kunlun Mts., Hohxil Mts., Tongtianhe plain, Tangula Mts., Cuonahu Lake as well as Damxung and Yangbajain basins, and three extremely unstable regions, Xidatan, western Gulu basin and Yangbajain basin, are identified along the Golmud-Lhasa Railway (Fig.23).

Acknowledgement: The First Academy of Railway Prospecting and Railway Design provided topographic maps at scale 1:2000 along the Golmud-Lhasa railway route for this study. Thank Professor Patrick J. Barosh for his careful review of this report.