John Keay "The Gilgit Game".
|
Turn back from the peaks of the Karakorams and face due south. Here lies the true horror of the Himalayas. This time there is no deep and distant perspective; the horizontal is unrepresented. You are staring at a wall; it rears from the abyss at your feet to a height for which the neck must crane back. Such is Nanga Parbat, ´the Naked Mountain´; its navel now confronts you. More a many peaked massif than a single mountain, Nanga Parbat marks the western extremity of the Great Himalaya; it is a buttress worthy of its role. John Keay "The Gilgit Game". |
Nanga Parbat, Urdu for Naked Mountain, is an 8000+ m peak on the northernmost edge of the Western Himalayan Syntaxis (Figure 2). The mountain, named for its southern face which is so steep it holds no snow, exhibits the world´s greatest continental relief, ~7000 m in 21 horizontal km. As you approach the mountain, a sign on the Karakoram Highway advises you to look toward the peak of ´Killer Mountain´, Nanga Parbat´s second name.
Figure 2. Regional and local setting. Area in pink on the regional map is the Kohistan-Ladakh island arc terrain. Red box outlines Nanga Parbat and is enlarged at right. The rocks of the massif are predominantly 1.85 Ga Pre-Cambrian gneisses of Indian crust exposed from beneath Asia. MKT=Main Karakoram Thrust, MMT=Main Mantle Thrust, MCT=Main Central Thrust, MBT=Main Boundary Thrust, STDS= Southern Tibetan Detachment System, ZSZ=Zanskar Shear Zone
Nanga Parbat, part of the Nanga Parbat-Haramosh massif, is an anomalous north-south extension of Indian crust into the Karakoram of Asia. The rocks of the massif are predominantly 1.85 Ga Pre-Cambrian gneisses (Zeitler et al., 1989). To the west and east lie mafic rocks of the Late Cretaceous and Eocene Kohistan-Ladak Island arc captured during the collision of India and Asia. The contact between the island arc and Indian plate rocks, the Main Mantle Thrust, is equivalent to the Indus suture in the central and eastern Himalaya.
The rocks at Nanga Parbat are unique in that they exhibit very young high-grade metamorphic and igneous activity, very high strain rates, and what can best be characterized as fierce and sustained exhumation at rates locally as high as 5 km/My for the last 3 My (Winslow et al., 1994; Zeitler et al., 1993). At Nanga Parbat, substantial reworking of the crust occurs today, although the initial collision between India and Asia occurred ~55 My ago. Perhaps even more remarkable is that at Nanga Parbat there is virtually no evidence for early Himalayan metamorphism. While these rocks were clearly involved in a major collisional event, recent processes have completely obliterated any igneous or metamorphic signature of the original collision.
The highest grade rocks at Nanga Parbat are highly migmatized and intruded by kilometer-sized granite plutons and granitoid dikes and veins. These intrusions yield ages as young as 1 Ma (Zeitler et al., 1993). Stable-isotope results from rocks in the core of the massif and in fault zones adjacent to the massif indicate that a two component hydrothermal system is active, a shallow level system dominated by fracture flow of meteoric waters and a deeper system involving circulation of magmatic or metamorphic fluids (Chamberlain et al., 1995). Active hot springs and evidence of recent hydrothermal activity are abundant within the massif and along its margins.
In order to characterize the active processes reworking the continental crust at Nanga Parbat we have embarked on a multidisciplinary study using techniques in geochronology, petrology, structural geology, geomorphology, geochemistry, and geophysics. Our primary objective is to assess the processes responsible for the severe tectonic and metamorphic overprinting observed at Nanga Parbat, processes that are usually taken as an indication of plate collision.
Extremely rapid exhumation, the presence of hot springs, young intrusive rocks and young metamorphism all suggest an anomalous thermal structure and perhaps even partial melt zones or magma lie beneath Nanga Parbat, making it an ideal candidate for seismic studies. In 1996, we deployed 60 IRIS PASSCAL instruments, (10 broadband and 50 short period stations), in NE Pakistan to record local and regional events. Our principal goal is to characterize the crustal structure and fault kinematics at Nanga Parbat. Ultimately, we hope to infer the thermal structure beneath the massif and to help constrain geodynamic models of uplift.
Deploying seismometers in the Himalayas of Pakistan is an interesting experience. Our deployment was the culmination of several years of logistical field work including many visits to various ministries in Islamabad, numerous cups of tea, scouting visits to the field area and a pilot project conducted in late summer/early fall of 1995. By the time our experiment was finished, the PASSCAL instruments had been transported by almost every means imaginable: airplane, Bedford truck, local jeep, porter´s back and donkey. The topography in the area is extreme (Figure 3). The majority of the array (56 of 60 stations) was deployed in an area roughly 60 x 60 km in and around the massif. Four broadband instruments were deployed in a more regional context to help locate regional events. Of the remaining six broadband instruments, three were deployed on the west side of the massif and three on the east. The short period stations at Nanga Parbat were deployed flanking the west and east sides of the massif along the Karakoram Highway and Astor river sections. Access to the interior of the massif was obtained along three primary glacial valleys, Bunar Das, Tato Valley, and Rupal Valley. Sites in these three valleys sit on Nanga Parbat gneisses. Reference sites, not on Indian crust, were deployed on the mafic rocks of the Kohistan-Ladak island arc. Station elevations ranged from 1100 - 4100 m. Station spacing varied between 1 and 10 km.
 |
Figure 3. Thematic mapper image with station locations. Stations, shown by red circles, were located along the Indus River section (NW side of Nanga Parbat Massif) and the Astor River section (NE side of Nanga Parbat Massif). Access to the interior of the massif was gained along three primary glacial valleys, Bunar Das (west side), Tato Valley (NS off the Indus river), and Rupal Valley (south side). Ice is blue. |
Our array recorded data for four months. We deployed our stations in Late May and early June as the snowline receded and we pulled them out at the end of September as the snowline advanced. This past year was an unusually late and wet winter. During our deployment, in May, we had to contend with an amazing array of landslides, rockfalls, high water, low snow, and active mud flows, all contributing to road blocks. Local jeeps ran shuttles between roadblocks. We moved our equipment by a series of leap frogging jeep rides, portering the equipment across the road block, then reloading in new jeeps to carry on. About half of our sites could be accessed fairly easily along roads and a short walk (Figure 4). The remaining sites could only be accessed by trekking (Figure 5).
 |  |
| Figure 4. Local road on the Astor River section. Road has 21 hairpin turns. Short period station was deployed 200m north of hairpin 19. | Figure 5. Makeshift bridge crossing glacial stream. |
 Figure 6. Typical site installation. |
At the beginning of our field season a loop up and back one of the glacial valleys took between 5 and 7 days. By the end of the field season, with our lungs and legs in better shape it took only 3-4 days. We found ample bedrock sites in the area and general interest in what we were doing and lots of help setting up sites and servicing the instruments (Figure 6). |
For the most part, our stations were sited away from villages. We worked in remote valleys at relatively high elevation and arranged for local shepherds to watch the stations between service runs. We serviced the stations on a three week interval. We rarely had problems receiving permission to site stations, although one person later changed his mind because he was worried the station might be damaged and he would be held responsible. When we arrived to service this station we found it was gone. The original chowkidar (watchman), had talked one of his neighbors into having the ´macheen´ located on his property. Together, they dug up and redeployed a short period station and did quite a respectable job. The L22 was oriented almost due north, was buried 12" deep in soil on bedrock, the solar panels were facing south ~30° from horizontal, and all the cables were properly connected. As it turns out, the powerboard had shorted during the move, so there was no station power, but other than that, the installation was great!
One of our more delicate negotiations took place after we installed a broadband station in Eid Gah near Astor Village. The day after our installation, a massive mudflow started. The mudflow ran for three days, washed out the only road servicing the village and nearly damned the Astor river. The local villagers were convinced that our station had triggered the flow. It took much talking, drinking of tea, and intervention from the head cleric to calm things down and, as we were temporarily stuck in the village, this made for a rather anxious few days. In another village, we arrived to service a broadband station and found that the local watchman had chained his (large!) dog to the site to serve as a watchdog.
As we begin to analyze our data we have identified some 2000 associated events. Primary source locations in the region include: the Pamirs and Hindu Kush to the northwest, the Karakoram and NE terminus of the Baluchistan arc to the north and northeast, the Himalayan arc to the southwest, the Hazara arc and Kashmir to the south and southeast. While we anticipated recording these regional events, prior to our survey little was known about the local seismicity at Nanga Parbat. A temporary regional array deployed as part of the Karakoram Project recorded only three events near the massif in a six week recording window (Yielding et al., 1984). However, our array recorded as many as 5-8 small magnitude local events per day (Figure 7).
Figure 7. Record section from local event (s-p=3.2s). Three small magnitude events were recorded within a single three minute triggered window. Event locates beneath Tato Village.
Preliminary locations indicate that local seismicity is restricted to very shallow depths consistent with high geothermal gradients and a shallow brittle-ductile transition as suggested by the petrologic observations. Some of the local events show evidence of shear wave splitting, presumably due to anisotropy associated with the metamorphic fabric of the rocks. We have also recorded events associated with hydrothermal activity at the Tato hot springs. Nanga Parbat is also very active in terms of avalanches and rockfalls and we recorded these as well, both on film and on our stations (Figure 8). Other sources of triggers were local goats. Even though we were in remote valleys at high elevation there were goat tracks everywhere.
 |  |
| Figure 8. Avalanche off north face of Nanga Parbat, and associated record section. | |
Strong thermal anomalies will result in seismic velocity and attenuation anomalies that we hope to map using seismic tomography. Even small percentages of partial melt (on order of 2%) cause a rapid decrease in P and S wave velocity, and attenuation increases by 3-4 orders of magnitude. While we have regional source locations from a variety of azimuths and abundant local events, by far the most abundant source is the Hindu Kush. Events from this region originate at 200-300 km depth approximately 200 km to the northeast and serve as a beam source to illuminate the structure beneath the massif. Another objective of our study is to use local seismicity to map the geometry and kinematics of active faults responsible for uplift at Nanga Parbat. While remarkably high denudation rates have exposed the Indian-plate rocks from beneath the over thrust Kohistan terrane, the actual uplift mechanism is not clear. As yet no obvious young extensional faults have been identified, so, at least for now, tectonic denudation as observed in other parts of the Himalaya does not seem a viable mechanism. The massif is bound on the west by the Raikot-Liachar fault which has been mapped as both a thrust fault and as a strike-slip fault. The fault is a young active feature that in certain places thrusts Pre-cambrian gneisses over glacial till. Surface mapping provides no constraints on the geometry of the fault at depth, and kinematic indicators in the fault zone itself are ambiguous in terms of a consistent sense of shear.
Our study at Nanga Parbat is still a work in progress. As far as we know, this is one of the densest deployments of seismometers in a active mountain belt. Our data set has the potential to look in detail at fault-slip behavior along a major crustal thrust fault as well as identify the presence or absence of partial melt zones in the crust beneath the massif. As we progress with our analysis, we have the great advantage of being able to help constrain our seismic observations with petrologic, geochronologic, structural, and magnetotelluric data. The ability to integrate multi-disciplinary data sets is valuable asset in this project.
