Seismotectonics: Himalayan, SCR and Reservoir-Induced Earthquakes
CSIR NET geophysics: Himalayan thrusts and great earthquakes, stable continental India seismicity, and reservoir-induced seismicity at Koyna.
- 20 min
On this page · 13 sections
Prerequisites
- Earthquake Sources, Moment Tensor and Focal Mechanisms — reading mechanisms and computing \( M_0 \) and \( M_w \).
- Earthquakes: Causes, Measurement and Seismicity — the Part B list of Indian earthquakes and the inter- versus intraplate contrast.
- Himalayan Orogeny and Supercontinents — the Himalayan tectonic units and thrusts from the geology side.
- Stress, Mohr Circle, Failure Criteria and Fault Reactivation — effective stress and the Coulomb criterion, the mechanics of triggering.
The Hook: An Earthquake Where None Was Expected
What Seismotectonics Uses
Seismotectonics combines several kinds of evidence to characterise active deformation:
| Evidence | What it gives |
|---|---|
| Seismicity catalogues (relocated hypocentres) | which structures are active; depth extent of the seismogenic layer; Wadati–Benioff zones |
| Focal mechanisms and moment tensors | style of faulting (thrust, normal, strike-slip); P and T axes; regional stress by stress inversion |
| Seismic moment summation | seismic strain rate; compare with plate rates to see what fraction of motion is seismic (Kostrov’s method) |
| Geodesy (GNSS, InSAR) | interseismic loading, locked versus creeping zones, coseismic slip |
| Active-fault geology and palaeoseismology | surface ruptures, trenching, dating of past earthquakes and recurrence |
| Historical records | intensities and damage from pre-instrumental earthquakes |
| Geophysical structure | crustal thickness, faults at depth, weak zones (from tomography, receiver functions, gravity, magnetotellurics) |
Seismic coupling is the fraction of plate motion released in earthquakes. If the moment released over a long period is less than the moment that would accumulate from full locking, either the fault creeps aseismically or a large earthquake is overdue. Slip deficit \( = \) (loading rate) × (time since last rupture) is the slip available for the next earthquake, if the fault is fully locked.
$$\boxed{\bar D_{\text{deficit}} = \dot s\,t}\qquad M_0 = \mu A\,\dot s\,t$$The Himalayan Arc
Structure
The Himalaya is the result of the continuing collision of India with Eurasia, which began in the Cenozoic after the closure of Neo-Tethys. From south to north, the major structures are:
| Structure | Character |
|---|---|
| Main Frontal Thrust (MFT) | youngest thrust; places Siwalik foreland sediments over the Indo-Gangetic alluvium; the surface trace where great earthquakes can rupture to the surface |
| Main Boundary Thrust (MBT) | thrusts Lesser Himalayan rocks over the Siwaliks |
| Main Central Thrust (MCT) | thrusts Higher (Greater) Himalayan crystallines over Lesser Himalayan rocks; inverted metamorphism |
| South Tibetan Detachment (STD) | north-dipping normal-sense shear zone above the Higher Himalaya |
| Indus–Tsangpo Suture Zone | the suture between Indian and Asian plates |
At depth, the MFT, MBT and MCT are thought to root into a gently north-dipping décollement, the Main Himalayan Thrust (MHT), on which the Indian plate underthrusts the Himalaya. The MHT is the source of the great Himalayan earthquakes.
The locked megathrust model
Geodesy and seismicity together give a simple picture:
graph LR
A["Indo-Gangetic plain
(foreland)"] --> B["MFT at surface"]
B --> C["Shallow MHT
LOCKED in interseismic period
width about 100 km"]
C --> D["Locking line
belt of microseismicity
beneath the Higher Himalaya front"]
D --> E["Deep MHT
creeps aseismically
beneath Tibet"]- The shallow part of the MHT, from the MFT northward for roughly 100 km, is locked between great earthquakes.
- North of the locking line, the deeper, hotter part of the MHT creeps. Stress concentrates at the transition, which is marked by a belt of moderate and small earthquakes roughly beneath the front of the Higher Himalaya.
- Great earthquakes nucleate near the locking line and rupture southward up-dip along the locked MHT. Some ruptures reach the surface at the MFT; others stop short (blind ruptures).
Mechanisms
- Thrust focal mechanisms with one shallow north-dipping nodal plane (the MHT) and P axes roughly perpendicular to the local arc dominate the front. The P-axis direction rotates along the arc, staying roughly perpendicular to it.
- Normal faulting with roughly E–W extension occurs in southern Tibet (N–S rifts), reflecting orogen-parallel extension of the plateau.
- Strike-slip faulting occurs in places, for example in the Sikkim region (2011 Sikkim earthquake, \( M_w \) about 6.9).
Great and major Himalayan earthquakes
| Earthquake | Size (approx.) | Note |
|---|---|---|
| 1897 Shillong (Assam) | \( M \) about 8 | on the Shillong Plateau, south of the Himalaya proper; widely interpreted as reverse faulting that uplifted the plateau as a pop-up block |
| 1905 Kangra | \( M \) about 7.8 | NW Himalaya |
| 1934 Bihar–Nepal | \( M \) about 8 | eastern Nepal; widespread liquefaction in the Gangetic plain |
| 1950 Assam (Assam–Tibet) | \( M_w \) about 8.6 | eastern syntaxis; one of the largest continental earthquakes recorded |
| 1991 Uttarkashi | \( M \) about 6.8 | Garhwal Himalaya |
| 1999 Chamoli | \( M \) about 6.6 | Garhwal Himalaya |
| 2005 Kashmir (Muzaffarabad) | \( M_w \) about 7.6 | NW Himalaya, Pakistan-administered Kashmir; very high death toll |
| 2015 Gorkha, Nepal | \( M_w \) about 7.8 | ruptured the MHT eastward beneath Kathmandu; did not reach the surface (blind) |
The stretch between the 1905 Kangra and 1934 Bihar–Nepal ruptures, without a great earthquake in the instrumental and recent historical record, is the central seismic gap (the 2015 Gorkha earthquake broke only its eastern part). Palaeoseismic trenching across the MFT has found evidence of large surface-rupturing earthquakes in past centuries, implying that some Himalayan earthquakes are larger than any recorded instrumentally.
Beyond the Himalayan front
India’s plate-boundary seismicity also includes the Hindu Kush (intermediate-depth earthquakes in a steep, nearly vertical slab), the Indo-Burman arc (intermediate-depth events in the eastward-dipping Indian slab), the Andaman–Sumatra subduction zone (the 26 December 2004 \( M_w \) 9.1–9.3 megathrust and tsunami) and the Chaman and other transform-boundary faults in the west.
Stable Continental Region (SCR) Earthquakes
Stable continental regions are the parts of continents away from active plate boundaries and without significant Cenozoic deformation: cratons, old orogens and rifted margins. Peninsular India is one of them.
Characteristics
| Feature | SCR earthquakes | Plate-boundary earthquakes |
|---|---|---|
| Rate | low; long and irregular recurrence (often thousands of years on one fault) | high; recurrence of decades to centuries |
| Strain rate | very low (GNSS often near detection limits) | high |
| Location | often on reactivated ancient structures: failed rifts (aulacogens), rifted margins, old sutures and shear zones | on the plate boundary |
| Prior seismicity | often none known before the event | usually well known |
| Depth | mostly upper to mid crust, some deeper | depends on boundary type |
| Attenuation | low (cold, old lithosphere), so shaking felt over large areas | higher |
| Stress | regional stress transmitted from plate boundaries (ridge push, collision), plus local concentrations | plate-boundary loading |
Why do they happen at all? Proposed explanations, not mutually exclusive:
- Weak zones in a strong lithosphere (rifts, sutures, intrusions) fail under the regional stress field, which in India is largely compressive from the India–Eurasia collision.
- Local stress concentrations: density contrasts, intersecting faults, flexure, erosion and unloading.
- Fluids lowering effective normal stress.
- Triggers: reservoir loading, mining, injection.
- Some SCR earthquakes may be part of clusters that migrate between faults over long times, making the last earthquake a poor guide to the next.
Indian SCR earthquakes
| Earthquake | Setting |
|---|---|
| 1819 Kachchh (Allah Bund) | Kachchh (Kutch) rift region; uplift of a ridge (Allah Bund) across a river channel |
| 1967 Koyna | reservoir-triggered, western Deccan Volcanic Province |
| 1993 Latur (Killari) | Deccan Traps of Maharashtra; about \( M_w \) 6.1–6.2; shallow reverse faulting in an area with little known earlier seismicity; very high casualties from heavy stone houses |
| 1997 Jabalpur | Narmada–Son lineament zone, a major ancient rift and shear zone |
| 2001 Bhuj | Kachchh rift basin; about \( M_w \) 7.7 reverse-faulting earthquake on a blind fault; one of the most damaging Indian earthquakes |
The Kachchh rift stands out: a Mesozoic failed rift now under compression and inverted, which has hosted two great intraplate earthquakes (1819 and 2001) within two centuries.
Reservoir-Induced (Triggered) Seismicity (RIS)
Definitions
- Induced seismicity: earthquakes caused by human activity: reservoir impoundment, mining, fluid injection (wastewater disposal, enhanced geothermal systems, hydraulic fracturing), and fluid extraction.
- Triggered is often preferred for reservoirs: the reservoir changes stress by a small amount (typically a few tenths of a MPa or less) on faults that were already close to failure; the tectonic stress supplies most of the energy.
Classic examples
| Reservoir | Note |
|---|---|
| Lake Mead (Hoover Dam), USA | among the first recognised cases of seismicity following impoundment (1930s) |
| Koyna (Shivajisagar), India | among the largest widely accepted cases: 1967, magnitude about 6.3; seismicity continuing for decades, later also linked with the nearby Warna reservoir |
| Kariba (Zambia–Zimbabwe), Kremasta (Greece), Xinfengjiang/Hsinfengkiang (China) | other cases with mainshocks above magnitude 6 in the 1960s |
Only a small fraction of large reservoirs are associated with significant seismicity. Deep reservoirs are more often associated with it, but depth alone does not decide.
Mechanisms
Using the Coulomb failure function with compressive normal stress positive,
$$\boxed{\Delta \mathrm{CFS} = \Delta\tau - \mu\,(\Delta\sigma_n - \Delta p)}$$where \( \Delta\tau \) is the change in shear stress in the slip direction, \( \Delta\sigma_n \) the change in normal stress, \( \Delta p \) the change in pore pressure and \( \mu \) the friction coefficient (often 0.6–0.8). A positive \( \Delta \mathrm{CFS} \) brings a fault closer to failure.
A reservoir acts in three ways:
- Elastic loading by the weight of water: changes \( \Delta\tau \) and \( \Delta\sigma_n \) immediately. It favours normal faulting beneath the reservoir (vertical load increases \( \sigma_1 \) for a normal-fault regime) and tends to stabilise thrust faults directly beneath (it clamps them).
- Undrained pore-pressure response: compression of the rock instantly raises pore pressure by \( \Delta p = B\,\Delta\sigma_{mean} \), where \( B \) is Skempton’s coefficient (0 to 1).
- Drained response by diffusion: water from the reservoir slowly raises pore pressure at depth along permeable faults and fractures. Diffusion causes a delay between filling and seismicity, scaling as
where \( L \) is the distance (or depth) and \( c \) the hydraulic diffusivity (m²/s). Values inferred from reservoir and injection seismicity are commonly in the range 0.1 to 10 m²/s.
Characteristics often reported for RIS
- Seismicity correlates with water level: rate of loading, the maximum level reached, the duration of high stands, and sometimes rapid drawdown.
- Initial (rapid) response: seismicity soon after first filling, attributed to elastic loading and undrained pore pressure.
- Protracted (delayed) response: seismicity that continues for years, linked to diffusion and seasonal loading cycles, as at Koyna–Warna.
- Higher b-values than regional tectonic seismicity, and sequences in which foreshocks are common (swarm-like behaviour).
- The magnitude of the largest aftershock is often closer to the mainshock than in typical tectonic sequences.
Memory Tricks
Don’t Confuse
| This | vs | That |
|---|---|---|
| MFT: youngest, southernmost, at the surface | MCT: older, crystalline over Lesser Himalaya | |
| MHT: the décollement at depth (source of great quakes) | STD: north-dipping normal-sense shear zone | |
| Locked shallow MHT (stores strain) | Creeping deep MHT (releases strain aseismically) | |
| 1897 Shillong Plateau (south of the Himalaya) | 1950 Assam (eastern syntaxis, larger) | |
| Interplate Himalayan thrusts | SCR Peninsular earthquakes on old rifts | |
| Induced: human activity supplies much of the stress | Triggered: human change is a small push on a fault near failure | |
| Undrained pore-pressure rise (instant) | Drained diffusion (delayed by \( L^2/c \)) | |
| Seismic coupling: fraction of slip in earthquakes | Locking: fault not slipping in the interseismic period |
Common Mistakes
Practice
All questions below are original practice questions written for this page. They are not from any CSIR NET paper.
Level 1: Foundation
- Name the décollement into which the main Himalayan thrusts are thought to root.
- Which Indian earthquake is the classic example of reservoir-triggered seismicity?
- Write the Coulomb failure function change in terms of \( \Delta\tau, \Delta\sigma_n, \Delta p \) and \( \mu \).
Answers
- The Main Himalayan Thrust (MHT).
- Koyna, 1967.
- \( \Delta\mathrm{CFS} = \Delta\tau - \mu(\Delta\sigma_n - \Delta p) \), compression positive.
Level 2: Part B Style
Practice Q1. Which ordering of Himalayan structures from south to north is correct?
(a) MCT, MBT, MFT, STD (b) MFT, MBT, MCT, STD (c) MBT, MFT, STD, MCT (d) STD, MCT, MBT, MFT
Answer: (b)
The thrusts young southward: MFT (frontal, youngest), MBT, MCT; the STD lies above the Higher Himalayan crystallines, north of the MCT.
Practice Q2. Which statement about stable continental region earthquakes is correct?
(a) They occur mainly on young faults formed in the present stress field. (b) Their recurrence intervals on a given fault are usually a few decades. (c) They often reactivate ancient rifts and sutures under far-field stress. (d) Because the lithosphere is cold, their shaking attenuates faster than at plate boundaries.
Answer: (c)
Reactivation of old weak zones (Kachchh rift, Narmada–Son zone) is the hallmark. Recurrence is long, and attenuation in old cold lithosphere is low, so shaking spreads farther.
Practice Q3. Fluid diffuses from a reservoir to a fault at 4 km depth through rock of hydraulic diffusivity 1 m²/s. The order of magnitude of the delay before the pore-pressure rise reaches the fault is:
(a) 2 days (b) 6 months (c) 5 years (d) 50 years
Answer: (b) 6 months
\( t \sim L^2/c = (4000)^2/1 = 1.6\times10^7 \) s \( = 1.6\times10^7/86400 \approx 185 \) days, about 6 months. Factor-of-a-few differences arise from the exact geometry of the diffusion solution, but the order of magnitude is the point.
Practice Q4. A strike-slip-dominated earthquake in the Himalayan collision zone would be most consistent with:
(a) slip on the shallow locked MHT (b) slip on transverse faults cutting across the arc, such as in Sikkim (c) normal faulting on the STD (d) a deep earthquake in the Indian slab at 600 km
Answer: (b)
MHT events are thrusts; the STD is normal-sense (and not the main source of large events); there is no 600 km slab beneath the Himalaya. Strike-slip events occur on transverse structures, as in the 2011 Sikkim earthquake.
Level 3: Part C Style
Practice Q5. A segment of the Himalayan front 150 km long has a locked width of 100 km and is loaded at 20 mm/yr. Its last full rupture was 600 years ago. Take \( \mu = 3\times10^{10} \) Pa. Which statements are correct?
- The slip deficit is 12 m.
- The accumulated moment deficit is about \( 5.4\times10^{21} \) N m.
- Releasing it in one earthquake would give \( M_w \) about 8.4.
- Releasing it instead as \( M_w \) 7.8 earthquakes would need about four such events.
(a) 1 and 2 only (b) 1, 2 and 3 only (c) 1, 2, 3 and 4 (d) 2 and 4 only
Answer: (b) 1, 2 and 3 only
- \( 0.020\times600 = 12 \) m. True.
- \( M_0 = 3\times10^{10}\times(1.5\times10^5\times1.0\times10^5)\times12 = 3\times10^{10}\times1.5\times10^{10}\times12 = 5.4\times10^{21} \) N m. True.
- \( M_w = \frac23(\log_{10}5.4\times10^{21} - 9.1) = \frac23(21.732 - 9.1) = 8.42 \). True.
- One \( M_w \) 7.8 event: \( M_0 = 10^{1.5\times7.8 + 9.1} = 10^{20.8} = 6.31\times10^{20} \) N m. Number \( = 5.4\times10^{21}/6.31\times10^{20} = 8.6 \): about nine events, not four. False.
The trap in statement 4 is to treat magnitude as if it were linear in moment. A difference of 0.6 units is a moment ratio of \( 10^{1.5\times0.6} = 10^{0.9} \approx 7.9 \) (here 8.6, because the single event is \( M_w \) 8.42, not exactly 8.4).
Practice Q6. Seismicity near a reservoir is monitored for six years after first filling. Observations: (i) a burst of small earthquakes within weeks of the first rapid filling, under the reservoir; (ii) from year 2 onwards, seismicity peaks each year about three months after the annual maximum water level, at depths of 4–8 km; (iii) the b-value of the sequence is 1.3, against 0.9 for regional tectonic seismicity; (iv) focal mechanisms under the reservoir are mostly normal faulting. Which interpretations are reasonable?
- Observation (i) is consistent with elastic loading and an undrained pore-pressure response.
- The delay in (ii) suggests pore-pressure diffusion, with a diffusivity of order \( L^2/t \approx (6\ \text{km})^2/(3\ \text{months}) \approx 5 \) m²/s.
- The higher b-value in (iii) is a commonly reported feature of reservoir-triggered sequences.
- Normal faulting in (iv) means the reservoir load must have stabilised the faults.
(a) 1, 2 and 3 only (b) 1 and 4 only (c) 2, 3 and 4 only (d) 1, 2, 3 and 4
Answer: (a) 1, 2 and 3 only
- (1) Immediate response = elastic load + undrained pore pressure. Reasonable.
- (2) \( c \approx (6000)^2/(90\times86400) = 3.6\times10^7/7.78\times10^6 = 4.6 \) m²/s, about 5 m²/s, inside the commonly reported 0.1–10 m²/s range. Reasonable.
- (3) Higher b-values are often reported for RIS. Reasonable.
- (4) Wrong: a vertical water load increases the vertical stress, which promotes normal faulting beneath the reservoir. It is thrust faults that a vertical load tends to clamp.
Practice Q7. Which statements about Himalayan seismotectonics are correct?
- Interseismic GPS shows the shallow part of the MHT to be largely locked.
- A belt of moderate seismicity roughly beneath the Higher Himalayan front marks the transition between the locked and creeping parts of the MHT.
- Normal faulting with roughly E–W extension occurs in southern Tibet.
- The 2015 Gorkha earthquake produced a large surface rupture along the MFT.
(a) 1, 2 and 3 only (b) 1 and 4 only (c) 2, 3 and 4 only (d) 1, 2, 3 and 4
Answer: (a) 1, 2 and 3 only
Statements 1–3 summarise the locked megathrust model and the extension of the plateau. Statement 4 is false: the Gorkha rupture did not reach the surface.
Quick Revision Table
| Idea | Fact or formula |
|---|---|
| Seismotectonic data | catalogues, mechanisms, moment summation, geodesy, palaeoseismology, history |
| Slip deficit | \( \dot s\,t \); \( M_0 = \mu A\dot s t \) |
| Himalayan thrusts (S → N) | MFT, MBT, MCT; STD normal-sense; all root into MHT |
| Convergence across Himalaya | roughly 15–20 mm/yr (GPS) |
| Locked MHT | about 100 km wide; microseismic belt at locking line |
| Mechanisms | thrust at front (P ⟂ arc); normal (E–W extension) in S Tibet; strike-slip on transverse faults |
| Great quakes | 1897 Shillong, 1905 Kangra, 1934 Bihar–Nepal, 1950 Assam (\( M_w \) about 8.6) |
| Recent | 2005 Kashmir (\( M_w \) about 7.6), 2015 Gorkha (\( M_w \) about 7.8, blind) |
| Seismic gap | central gap between 1905 and 1934 ruptures |
| SCR features | low rate, long recurrence, reactivated rifts/sutures, low attenuation |
| Indian SCR | 1819 Kachchh, 1967 Koyna, 1993 Latur, 1997 Jabalpur, 2001 Bhuj (\( M_w \) about 7.7) |
| Coulomb | \( \Delta\mathrm{CFS} = \Delta\tau - \mu(\Delta\sigma_n - \Delta p) \); triggering at about 0.01–0.1 MPa |
| Undrained | \( \Delta p = B\,\Delta\sigma_{mean} \) (Skempton) |
| Diffusion delay | \( t \sim L^2/c \); \( c \) about 0.1–10 m²/s |
| RIS traits | water-level correlation; initial and protracted responses; higher b; foreshocks |
Key Takeaways
What’s Next
- Seismic Hazard and Earthquake Prediction: turning sources, rates and ground-motion models into hazard estimates.
- Travel-Time Residuals and Tomography: how the slabs, plumes and crustal roots behind seismotectonics are imaged.
- Structural Geology and Geotectonics: the rock record of the same deformation.
- Environmental Earth Sciences: earthquakes among the natural hazards in the Part B syllabus.