Sunday, August 18, 2013

Threshold Landscapes and Uttarakhand Devastation. Things we Must get Right Before we Reconstruct.

Preamble
 
The path to holy Himalayan places using the old pilgrim’s route (see figure bottom left) by foot was used not too long ago.  It served a soul-fulfilling purpose of the aam aadmi who carried their possessions in a cloth bundle on their head. They had no footwear, no warm clothes except perhaps for a chaddar of the coarsest (most comfortable and wieldy) kind; the older of them carried perhaps a stick. In my early perambulations (I don’t know if this is the right word … it just came to my mind) of the Himalayas during my non-formative chipko phase, I had walked around Vimlaji and Sunderlalji Bahuguna’s ashram in Silyara to Ghansali and Ghuttu and Gangi and I had looked wistfully at the pilgrims progressing towards Trijugi Narayan or Budha Kedar. They give substance to the image of holiness of a pilgrim’s journey. 
 
These are our people from our perennial  “हम तुम युग युग से ये गीत मिलन के, गाते रहे हैं, गाते रहेंगे” dreams. 
 
They are very different from the middle-class aam aadmi, say, of the Kejriwal/Anna-Hazare kind, who go in hordes and demand well-paved roads and other middle-class luxuries. 

Now, with our multimedia vision, the style of the journey has been encouraged to become the substance. There is a con in it --- the money-make-onics that goes in the garb of economic upliftment of the “poor” hill people. It is this “development” that seems to have brought down the Himalayas to this grim state of devastation which increases each passing year. The old pilgrim route (see figure above left) did not plunder, as they do now, the sides of the deep gorges of rivers for used by roads for buses. 
 
It is turning out that the Himalayas will not take the load and their accompanying paraphernalia. In this context, I quite like Anant Shankar’s latest cartoon on pili-grims (above right) which should be there in one of his Crazy Desi Books (see http://www.timeoutbengaluru.net/books/features/purr-desi). 
 
This blog is to make some science-based arguments on why we should leave the Himalayas alone and not attempt a reconstruction of the pilgrims’/soldiers’ route without truly understanding Her Majesty, Gaia or what we call Shakti. 
 
Most of my arguments may be re-stating older arguments, at best in a new way. Some may be just new for newness’ sake. I hope some will create debate and thereby contribute to a better understanding.
 
If traditional beliefs are based on experience, we should follow tradition. We should walk. In the last part of this blog I have given a solution for the pilrgrims’ jopurney which need not be entirely togue-in-cheek. 

Figuring out the Kedarnath Devasation of 16th-17th June 2013.

When starting this series of blogs just after the Kedarnath devastation of 17th June 2013, I planned to write on what I had observed on google map on that day thinking that part of the answers may have been written in the topology of the Himalayas. After looking at the features I planned writing on how temporary lakes that could have formed uphill because of landslides could have collapsed due to excessive rainfall and soaking of sandy/clayish soil so that there could be massive mudslides. I thought my knowledge was severely inadequate and I started reading up on the Himalayan disasters and instabilities. Within no time, however, I found that the lake-burst theory (see David Petley’s Landslide Blog, http://blogs.agu.org/landslideblog/) has been documented in some way. As recently noted in the Guardian “The origin of the disaster is beyond dispute: a glacier ruptured under the pressure of water from a severe cloudburst, raining tonnes of ice, water and rock on the Hindu pilgrimage town of Kedarnath, on the left bank of the Mandakini. The lake-burst theories have resulted in a dramatic video reconstruction of what could have taken place (http://in.screen.yahoo.com/headlines-today-recreates-flooding-mandakini-032414394.html).  

One of the first features that struck me when I looked at the google map using Wikimapia is the round doughnut shaped feature (Fig 1, left, circled in orange). This feature been indicated as Gandhi Sarovar because it seemed that part of Mahatma Gandhi’s ashes were immersed there. This place has also been assigned as Chorabari tal (lake) because it is on top of the Chorabari glacier. However, there seems to be another Chorabari tal indicated by a blue circle. When I looked at the same region through blink-wikimapia much of the snow cover was reduced (Fig 1, right). The feature inside the green circle is the area analyzed by David Petley.

In Petley’s blog we have (Fig 2) google earth pictures a few days before and a few days aafter the Kedaranath distaster on the morning of 16th May 2013. The main changes seen after the devastation (Fig 2 right) is shown by arrows. The snow or ice cover from the lower slopes seems to have gone. On the upper slopes the ice/snow cover rhas remained roughly the same if not actually increased. It suggests that the snow/ice cover was washed down by rain in the lower ranges and that it could have snowed in the higher ranges. The removal of snow seem to be quite widespread in the lower ranges. Snow melt or snow slide could have taken place all over the hills!

 

After a series of speculations in his Landslide blogs on the Kedarnath disaster, David Petley settled down (Reconstructing the events at Kedarnath using data, images and eye-witness reports, of 4th July 2013)  on a dam burst explanation that is outlined in Fig 3 left. Number 1 on this image is fresh muddy sediment which blocked water to build up a pool. Thre was a catastrophic breach at 2 of the blocked basin. The first flow from this breach was at 3 going south-east (north-south is in the top-bottom direction) and a relatively small “trickle” going southward towards Kedarnath at 5 using an older channel. The major breach was at 4 where, according to Petley, there is “substantial erosion” and “source of many of the boulders now seen in the town.” An earlier (2010) Bling map of the area is shown in Fig 3, right. A comparison of the two maps gives an idea of the extent of new breaches and new sand/mud covers to the north of Kedarnath.

This blog of Petley has a description by one constable, Kuldeep Singh Mehra, in which he said everything was over in a few seconds. On the morning of 17th June, Mehra came out to see a piling up of Boulders, rocks, sand and gravel around the temple complex. “Then I heard a piercing sound, like a storm hurtling towards us,” said Mehra. He ran back to the temple.“I screamed—run, run, run, get inside—and just as I entered the mandir, a huge wave threw me inside,” he says. “I was lifted so high that I was near the ceiling, and I grabbed the electric cables and held on.” He looked down to see heads bobbing up and down in the water, mouths twisted in screams, hands desperately reaching out for suspended bells, wires, anything. The water burst through a side door and carried away some pilgrims. Sand poured in, collecting 4-5ft deep. … The buildings that stood were packed with sand and gravel to the ceiling.“I knew immediately that no one could have survived,” said Mehra …“Everywhere we looked there were dead bodies,” said Mehra. “Lifeless hands and legs stuck out of the sand-packed windows and doors. One had red bangles on it, the kind you wear when you are newly married. … The Mandakini raged all around them, creating three channels that coursed through the temple town. (see http://www.livemint.com/Politics/JvCO1EbT7O7QfSV7BVUtPM/In-Uttarakhand-a-pilgrimage-of-horrors.html for more details)

It rained incessantly between the nights of 15th and 16th June. It rained more than 30 cm during this time. It is possible that unmelted snow blocks as well as the hill slopes slid and built temporary dams which breached to release the water. A large temporary lake could have formed at what has been called in some sites as the Chorabari tal or Gandhi Sarovar (circled in orange in Fig 1, left). This feature when empty could be (Fig 4, right, the features of the dried lake is seen) about 500 m in diameter, This dry reservoir tilts about 200 m from top to bottom and has walls about 100 m high (Fig 4, left). This gives a volume of about 10 million m3. If the whole area of the Kedarnath town (~ 200 m x 600 m) was instantly filled by this volume of water, the water level would require to be about 80 m high. That is too high. It is however consistent with a tsunami kind of wave passing across in a few seconds with heights of a few meters (say 5-10 meters). The “chorabari tal” is about 3km north and about 600-800 m higher so that water could have gushed down and hit Kedarnath with considerable speed. If this volume of water was now transferred to a river gorge say, 50-100 metres wide without dissipation there would be a cigar-shaped tsunami about 50-100 metres high if it is one kilometer long. This would have swept away all settlements along the Mandakini river for some time. There would have been other tsunamis that in other parts that would have added other “cigar-tsunamis”!


There does not seem to be many boulders that hit Kedarnath in this tsunami event. There was however a lot of sand, It has been said and seen in TV images that easily two-five metres of sand or soil were deposited along the river bed. That could make the soil content of the tsunami to be roughly even as large as 20 percent of the volume of water. That is a lot of sand all of which is carried down-stream to be deposited in the dams that are built. This is also a frightening thought in a way, even if it just a thought for the present. For example, if there was 30 cm rain and 10-20 percent of the rain-wash is sand or soil, then an average of 3-6 cm thick soil was scraped off the slopes in that day. One could think that this is a trivial amount. It is not so if one thinks that one-or two years of Himalayan tectonic movement has been washed off in a single day! The estimated amount of water released from the Gandhi Sarovar is 0.01 km3 which is about 1/200 the volume of the Tehri dam reservoir when full. If 5-10% of the sarovar volume is the amount of sand released from that one source, it will take 100 such “dam bursts” from natural reservoirs around 50 years to fill the Tehri dam to full capacity with sand. This is not a long time! On the other hand, it will only require to fill to the levels of the T3 T4 tunnels to cripple the dams (see later). This level is about 200 m below the top level of the dam. Since the reservoir itself is V-shaped the extent of sand required to fill and damage the dams would be much less,

Dam bursts are, of course, not the only reason for excessive flooding. Excessive rain in May before and the accompanying snow-slides and landslides over large parts of the Himalayan slopes could also give rise to flooding. One could argue that this year (2013) there was just too much rain.
The damaging excess rains may come anytime. For example in September of 2010, in the Almora district of Kumaon Himalayas nearly 30 cm of rain fell over two days after a few days of heavy rain that had already saturated the land with water. The rainfall density here was perhaps half that in the recent Uttarakhand devastation. The result of devastation was perhaps a quarter less in Almora with “… a swarm of landslides and debris flows, and a spate of surface water runoff and mobilised debris that swept, downslope, into the river network, creating a major flood surge in the region’s main drainage channel, the River Kosi (Kaushilya Ganga). River discharge rose from its pre-Monsoon level of 0.07m³/ sec (20th June 2010) to a peak of 618.1 m³/sec (18th September, 2010).” (from Landslide causes: Human impacts on a Himalayan landslide swarm, Haigh and Rawat, Belgeo 2011).


If one considers the whole of the Himalayas and the area of drainage to the Indo-Gangetic basin then the total sand/clay drained to the basin would be several km3 per year! Considering that the volume of soil in the Bengal fan (see previous blog) is nearly 60 million km3 one could argue that such floods have been happening once every year for the last 60 million years! Global warming would not be a unique cause.

The question that can now be asked is whether these devastations were natural disasters by themselves? Or, were the disaster accentuated by human interference that is compounded by human ignorance?

There are several institutes that have been set up in the Himalayas and around for the study of earthquakes and geology disaster management. One would have thought that after so many years one would have formed an independent-minded body of learning that is authoritative and respected by environment.

Maybe we did not learn enough especially when we may not have cared to learn enough. Maybe the powerhouses of learning --- the idealized casteless Brahmin class --- became subservient to the warrior/business class. This could have happened, say, once the Brahmins (in the east and the south of what is now India) who were Buddha’s disciples lost their sway to the Vedic Brahmins from the west.

Himalayas as Piles of Sand.

It is the sandy nature of the Himalayas that probably requires the most understanding, if our civil engineers are to execute their Himalayan projects with utmost professionalism. These slopes are sandy. They are threshold slopes which are critically balanced. The extent of sandiness, if that is the word, is not something that is reliably known. Science is just starting to understand on important aspect that is easy to understand and diagnose from simple geometrical aspects.

One of the scientific aspects in the context of Himalayan landslides must be the sandpile model (see previous blog). The basic premises of this model is simple to see. One may start off by looking at a sand particle as a ball. By itself the ball is stable on a flat land even if easily movable. If another ball is placed on top the structure is unstable and the ball will fall off. On the other hand when there are number of balls on a plane, and the lateral mobility is restricted, there is a possibility that there would be, say, a triangle of balls placed such that another ball may be placed on top of the triangle to form a more stable three-dimensional structure. The pile of balls or of grains of sand or sandpiles could start from this point on. As one adds more sand, more such piles would be formed, some of them bigger than the other. As the pile becomes bigger on adding more and more sand, there could two possibilities. In the first entropy (disorder) consuming ordered possibility, One could, in principle, make an ordered sandpile if one took the trouble of piling uniformly sized grains of sand in a close-packed (CP) manner just as they pack, say, atoms in metallic solids. The volume of close-packed fraction is 0.74. One could fill all voids and interstices by smaller particles as they do, say, in models for complex inorganic salts. In this perfectly ordered state the world could end up as a giant crystal.

The more complex possibility is that sand is piled up randomly without giving it time to organize to a more stable structure. An extreme case is that in which all the particle are randomly stacked on top of each other. The empirical finding with spherical objects randomly stacked (shaken and settled in oil, kneaded into balloons), is that the maximum density of random close-packed fraction is 0.63. Random close-packing (RCP) has high entropy and because of this it is unstable with respect to transition to a more densely packed CP ordered state. Sandpiles have this excess entropy or disorder and have a tendency to re-organize irreversibly to a more stable structure.

In the sandpile model one piles sand on sand to and the pile reorganizes as best as it can to a stable structure, all the time retaining its intrinsic disorder. A point is reached when the sandpile becomes a single uniform cone with a cone angle. As more and more sand is piled the pile becomes higher resisting the tendency to increase is base by spreading out and decreasing its slope angle. At one point, slope of the pile of sand increases to such an extent that the addition of more sand causes instability over a long range and the sandpile reorganizes. It reaches a critical state when an added grain of sand could cause more grains to fall which in turn cause others to fall and slide down the pile in some sort of a chain of slides which finally becomes an avalanche in some scale. There is a self-organization in this critical state (the self-organized critical state or SOC). The critical slope at this SOC state is close to 35o.

In this sandpile model one starts by piling up sand from zero. The other approach is to start with very large slope of a hard rock-like slab. One starts crushing the rock and looks at the way the sand or rubble piling up at the foot of the slope. The slope should remain constant once it reaches the critical value of ~ 35o. Evidence for this being the case was given in the previous blog. The fact that the slope is close to 35o does not mean that the slopes can be treated as sand as found on the beach. It probably means that when there is a critical state of a landslide the composite of rocks and boulders and clay and soil may be treated as a pile of sand in its critical state. The size of bigger rocks or wetness of clays does not seem to matter when there is a spontaneous (self-organised) landslide. When one disturbs the critical state by some activities, human or natural, there will be a reorganization to reach that critical state, with its accompanying disturbances or devastations depending on the scale of criticality.

In a recent publication Larsen and Montgomery have looked at de-classified spy satellite terrain maps around the region of Namche Barwa in the Tibet Plateau on top of Arunachal Pradesh (see Fig 5, bottom right). This region is prone to landslides. From the volume of landslides measured at various slopes they found (see Fig I, right; I will use Roman numerals for the main text, and Arabic numerals for the supporting write-up). In particular they examined whether relief in mountain ranges will rise as the uplift rate increases. There seems to be a threshold hillslope of ~ 35o. (see also comments on Figs 1, 2, 4 and Fig 5, right in supporting write-up). Above this slope the changes in slope due to tectonic uplift and incision by streams is balanced by landslides. Such a balance presumes that the crust of the Himalayas is close to being treated as sand-like loose rocks (in the Himalayan scale) due to severe weathering of rocks. Larsen and Montgomery demonstrate this hypothesis in the region around Namche Barwa where hillslopes do not vary by more than 10% even if rate of uplift and stream power can vary substantially. The angle of 35o is the critical angle for Per Bak’s self-organized sandpiles.   


It does not mean that once the critical angle is reached, the sandpile will be stable, say, towards road building or dam reconstruction. Robert Jackson’s blog on Sandcastles and Sandpiles (see http://robertjackson.info/index/2011/04/sandcastles-and-sandpiles-levi-on-entropy/) gives useful insight.

Take a pile of sand. It can be said that this object has a high degree of entropy. … it’s elements have an equal probability of being located anywhere in its system. In other words, … many different types of forces can impact on the pile and sand, and it will retain the same structure. Or to put it another way, the billion or so sand particles in the pile of sand, can be rearranged in many different ways and still retain the basic structure; a pile….of sand.

Basically it means that slopes formed from landslides on threshold slopes are unstable all the time. It means that the Himalayas form part of, what are called, threshold landscapes. Such landscapes are vulnerable to very long-range reorganizations (landslips?) that may be brought about by very small disturbances. Faced with these new data it is apparent that serious --- excludes headline seekers --- environmentalists have to do their homework again and again before convincing the nexus of temple head-priests and road contractors. The anti-dam, anti-road logic must emerge as out-right winners before more extreme disaster scenarios become more imminent if road-building and dam-building continues even at the present pace.

It is perhaps helpful to a layman to get a perspective on how the Himalayas were formed. After getting all the geology and all the tectonics and all the orogeny (forces and events leading to a large structural deformation of the Earth's crust and uppermost mantle due to the engagement of tectonic plates that results in the formation of long tracts of highly deformed rock --- from Wikipedia) right, one still requires a simple picture to start a model in one’s mind. Unfortunately, being simple is not sufficient, but if one is lucky it could mark a beginning. Part of a current picture for the formatin of the Himalayas is given in Fig 12 of the previous blog.

For some reason that I am not fully familiar with, one starts zillions of years ago with Gondwana land and the Eurasian plate to the north with a narrow. elongated Tethys Sea in between. Due to some event in the Cenozoic period (~ 60 million years ago) a plate which eventually became the Indian plate separated and collided with the northernplate taking a part of the Tethys sea along with it. The Tethys sea was filled with sedimentary rocks. The Indian plate subducted under the Eurasian plate giving rise to the upliftment of the present Tibetan plateau and the formation of the Himalayas.

The rocks in the Garhwal regions are young being between the neoproterozoic and the cretaceous periods. They are considered to be very young, being less than 1.0 billion year of age (paleoprotrozoic period is 2.5 to 1.5 billion years, neoproterozoic 1 bn – 540 mn years; Cambrian 540-480 mn years, Cretaceous 245-65 mn years; Cenozoic, less than 65 million years). Due to the pressure and heat generated by the collision the young sedimentary rocks undergo metamorphosis to resemble metamorphic rocks. What seems to be important is that the crustal rocks seem to become younger as one goes higher northwards.

The rocks here, such as the schists, slates, gneiss, dolomites, are not really homogenous over large length scales, but vary even in kilometer length scales or smaller (see Figs 6 and 7 of previous blog). These are also very flaky or layered with the plane of these layers being thought to be perpendicular to the direction of thrust. The rocks may therefore be considered to be stratified.


A simplified picture of the tectonic uplift that I have formed in my mind will probably go like this to begin with (this picture keeps changing with extent of information, but has been stable so far as thewritingof this blog is concerned). One starts (Fig 6, left) with alternating layers of soil-mantle and bedrock for, say, the Tibetan plate. For convenience we shall call the bedrock layer as the “rock” layer. The soil-mantle or the weaker rock layer in stratified layers of different rock types will be called as just “soil” layer. The uplift of this plate due to a subduction of the Indian plate takes place from left to right. The uplift is schematically sketched in the left of Fig 6, when there is no change in the nature of “soil” and rock layers with uplift. Above a critical angle of uplift there would be landslides within the “soil” layer In the right of Fig IV The rock layers are rendered more fragile and fracture and fall into the “soil” layers from which have been emptied either by landslides or erosion. The frequency with which this happens would increase when the slope increases above ~35o either by the tectonic uplift or because of erosion by fast water currents. Later, this “soil” which has been thus exhumed flows down into the flatter planes or through its rivers to the ocean and contributes to the fans in the ocean. At high uplift angles there will be little “soil” left and the rock layers would collapse over one another. One expects that due to drainage of the “soil” the height of the uplift can reduce as long as the rate of uplift is slower than the rate of “soil” exhumation events. We have shown this process only for the top few layers in Fig 6 without considering any reduction in height

On the basis of the above simple (not necessarily simplistic) picture, one gets the feeling that one may not get any insight that will help one to make, in the long run, safe projects for transport (pilgrims included), irrigation and hydro-electric projects. Every feature of the Himalayas is so fragile. Per Bak called his sand/pile model a self-organized critical state and thought that such a state is characteristic of living systems. One should then actually say that the Himalayas are very alive because it’s very sensitive sand pile state.

What is very important is that the association of sandpiles theories with mountain-scapes has been made much after decisions on Himalayan ecologies have been formed and transgressed.

An Oxford-Indian Contribuition

It is perhaps of some significance that the first papers which looked at sandpiles in Himalays appeared early in 1993 (Noever, Himalayan Sandpiles, Phys. Rev. E. vol 47) six years after Bak’s paper. This Noever paper was more concerned with the predictions of the frequency vs size aspects of the landslides in the Himalayas in the sandpile model.

The analyses in the paper by Noever was based on data published earlier in an Indian journal (Current Science, 1988) as ‘Entropy Minimizing Landslide Systems” by Haigh from Oxford, and Bartarya  and Rawat from Kumaon University. It is common practice to assume that landslide are related to environmental factors such as the geological morphology, the physical characteristics of rocks (lithology), the nature of the bedrocks and the extent of land cover. Haigh and coworkers followed the method of statistical analyses of landslides pioneered by Carrara of Italy that recognized that there is a linear rank-size rule (the plot of logarithm of volume of outfalls vs their rank in their size is linear) for landslide size distributions.

The paper by Haigh et al begins with the following:- “Landslides are complex dissipative systems which self-create on hill slopes provided with excess relief energy. This excess may be generated by processes which increase stress in, or which decrease the strength of, the hillslope. The result is the same. The hillside is transformed by the emergence of a landslide structure.” These few sentences have a considerable amount of language used by Bak (of sandpile) and correctly describes landslides as dissipative systems : “A system may import so much energy that it dissipate more entropy to the environment than it creates” They also included the fractal description: The landslide outfalls  can be called self-similar (fractal dimension: about 1.6). This is consistent with the existence of a single, maximum entropy attractor in the landslide system.” These descriptions are similar to those that are described by sandpiles which are also dissipative systems.

The major scientific interest of Haigh et al in their 1988 paper just cited, seemd to stem from their observation that while “… there are not only landslide system whose activities diminish from year to year there are the “chronic landslides” recognised by road engineers. Their landslide actvities increase from year to year.” They sought to connect the latter with the concept of “autopoiesis” using the language of evolution theory. Autopoiesis (self-producing and self-organizing) “… recognizes that as dynamic systems evolve they insulate themselves from fluctuations in their environment.” In dynamic system terms, the autopietic  landslides function to some degree as self-referential such that their current size and growth rate is essentially, a function of their own internal structures. Thus, in “chronic landslides” entropy is minimized there is a large independence from environmental control and it is “… identifiable through an evaluation of their own, internal, energy-generating, structures. In this case, the key structure seems to be the back wall created by a rotational slump, a feature which may be steeper and less stable than the hill side it replaces.
Later in 1993 Haigh (“Garhwal Himalaya: Ecology and Environment,”  edited by G. S. Rajwar) commented that the problem of detecting the geoecological correlates of landslide activity has proved sufficiently intractable that “some of the team” have become inclined to the view that the road-induced landslides are, essentially, independent of environmental controls. That is they are autopietic.
Part of this intractability is reflected in the report in 2011 cited above (“Landslide causes: Human impacts on a Himalayan landslide swarm” by Haigh and Rawat, Belgeo, 2011). Thus they write “apart from steepness of the hillslope and the height of the initial roadcut, few environmental variables correlated with enhanced landslide debris production. Indeed, statistical analysis indicates that the traditional suspects – deforestation, building construction, the collapse of agricultural terraces, the collapse of roadcut retaining walls, and the presence of other roads upslope of the landslides all tended to be associated with smaller than average outfalls. … The conclusion … was that neither human development nor local environmental factors played a significant role in the enhanced debris production of September 2010 … .

A revealing part of the report by Haigh and Rawat in their report is that on the long term effects. They had earlier studies the landslides on the Kilbury road of Nainital going through reserve forest and the Lower Mall of Almora that went through human habitation. In 1985, a 7.4 km stretch of the Kilbury road had 153 landslides while Lowe Mall had 88 landslides. In 2010, the Kilbury road had only 9 landslides while the Lower Mall had 108! The result is in agreement with the expectations of environmentalist that human activities and/or deforestation leads to deterioration.

For instance a well-cited report in 2001 on “The variability of root cohesion as an influence on shallow landslide susceptibility in the Oregon Coast Range” published in the Canadian Geotechnical Journal, found that old unharvested forests had four to five time better lateral root cohesion than even 123 year old industrial forests. Conversion of old unharvested forests to industrial forests increase rates of landslides. Further. the practice of using herbicides to remove weeds that are harmful to industrial forests further decreases root cohesion and increases the risk of landslide hazards.

This is what pioneer tree-hugging (chipko) environmentalists such as Chandiprasad Bhatt and Sunderlal Bahuguna had been shouting from roof tops.


In the article by Wu there is a plot (Fig 7) of events in a spatial vs temporal scale. Among other things one finds that to make good soil takes bigger (by nearly two order of magnitude) spatial (in km2) and temporal scales (in years) than that required for soil erosion. The temporal scale of 100 years is the time to get wisdom from several generations of wisdom. Industrial activity for profits cannot afford to wait wisely. Jairam Ramesh may have found that out when he was sent out o the environment ministry by his Prime Minister. 

Importance of Hierarchical Descriptions

I have a suspicion that much of the theoretical initiative on this work on Landslides was from Martin M Haigh of Oxford. He seems to have insights (his article on “Holons, fractals and the dynamics of evolving hierarchical systems” in 1989 in a Czechoslavakian book) on fractals and evolution and hierarchical systems as well as holons.  He seemed to have used the work of  T.  F. H. Allen and coworkers (Hierarchy : Perspectives for Ecological Complexity, 1982; Interlevel relations in ecological research and management: Some working principles from hierarchy theory, 1984).

Entities that belong to hierarchical systems have been called by Koestler as holons  (Greek holos = whole and the suffix on suggests a particle or part as in proton, neutron). Thus far, the term holon was familiar to me only in the context of resonating valence bond theory of condensed matter physics where the “hole” implies an absence of electron. My recent education on the hierarchical  “holon” starts from an article by J. Wu (Canadian Journal of Remote Sensing Vol. 25,, 1999, pp. 367-380) on “Hiearchy and Scaling: Extrapolating Information along a scaling ladder”. In this article Wu writes,

A hierarchical system has both vertical structure that is composed of levels and horizontal structure that consists of holons … . Hierarchical levels are separated, fundamentally, by characteristically different process rates (e.g., behavioral frequencies, relaxation time, cycle time, or response time). The boundaries between levels and holons are termed surfaces (Allen, 1982), which in space are the places exhibiting the highest variability in the strength of interactions (Allen et al., 1984). Surfaces filter the flows of matter, energy, and information crossing them, and thus can also be perceived as filters. … In hierarchical systems, higher levels are characterized by slower and larger entities (or low-frequency events) whereas lower levels by faster and smaller entities (or high-frequency events). … … On the other hand, the relationships between subsystems (holons) at each level are symmetric, and can be distinguished by the degree of interactions among components. That is, components interact more strongly or more frequently within than between subsystems or surfaces. For example, the strength of interactions between subatomic components is stronger than that between atoms which is in turn stronger than that between molecules… .”

In every level of such hierarchical systems the functional units may be thought to be double-faced as far as a one-dimensional hierarchy of levels is concerned. Looking down the hierarchy they seem to be whole dissembling to parts while going up the hierarchy they seem to be parts constituting the whole. Because of this double-faced aspect, holons are sometimes called Janus–faced. 

Of course, being Indians we do not like one-dimensional choice-less paths. We basically thrive in being anarchists as the speaker of our parliament seems to have finally recognized and which our opposition accepted as their birthright with the proviso that they be not called as such. If we are of the Buddhist or enlightened kind or even if we are of the Rama or non-philosophic kind, we can assume that the word Janus comes from yāna (spiritual journey starting, in Roman calendars, from January; when else?) where there is a pitryãna (path of the fathers) and ekayãna (destination). This would be a hierarchical system unlike, perhaps the pre-ordained devyãna (the path of the gods). The Buddhist would actually have Yāna as a vehicle to carry one though life using various (say, ten-fold) paths so that the yãna need not be one–dimensional with a forefather and a destination, although it may look one-dimensional when one is in one of the many paths.

Reverting back to landslides, the landslide landscape could have many paths of different complexities. It would then be a folly to treat all landslides on an equal footage. Thus to adopt methods first used in Calabria of Italy for use in the Garhwal or the Kumaon regions of the Himalayas. It would also be a mistake to think that the hierarchy of land-sliding events could be the same. One may think of landslides as part of natural processes as entropy maximizing. There are constraints that are holding the land together in an unstable state. Removing some bottlenecks or asperities will lead to a free flow of land till another bottleneck contains the flow. This is like letting water out of a bottle, into a sink into the drain into the gutter, into the river and into the ocean. This would be a slow process. But if we are to imagine that the ocean is allowed to come into the room the results would be a catastrophic.

All this depends on the nature of hierarchy. The problem of constrained disorder is there in the problem of gassy solids. The temporal relaxation of the hierarchy of constraints is sometimes described by a stretched exponential relaxation.  Specifically, the relaxation of a property, q(t), relative to its equilibrium value, q0, is given by a stretched exponential decay rate for a given temperature by q(t) = q0exp[-(t/te)b]   for a given characteristic time te and 0 < b < 1 (see previous blog “Himalayan Instabilities: What One Wanted to Ask but was Afraid to Know”, section 2c) When b = 1 there is the usual exponential decay of non-hierarchical systems.  

Instead of using a forward path, with time, t, always having a positive value going towards the future, one can go backward in time (towards the past) by choosing the time, t, to be before and after a fixed time, tfix such that (tfixt) could be positive or negative. If q(tfix) is the value of the time-dependent property, q(t), at t = tfix, by q(t) = tfixexp(-½(tfixt)½/te)b. The inverse value of the exponent, b, or 1/b, is a measure of the dimensionality of the system or heterogeneity (the number of parameters required to describe the system).

The property q(t) then gives the changes in property before and after a fixed time, tfix. The problem, here is that for complex systems or natural system, the decision of when a time is before tfix would require a premonition of tfix, which is not always available to us. If b and te is known analytically or empirically one could anticipate tfix for t < tfix.
In the case of earthquakes, for example, one may consider the main earthquake event to occur at a time tmax. The earthquakes before tmax would be foreshocks and those after tmax would be aftershocks. It turns out (see Fig 3 of previous blog) that the description of the foreshocks and aftershocks using the using inverse, R-1, of a renormalized length, R, which is a distance from the mainshock, may be fitted using this stretched exponential with t­fix º tmax.

This exercise helps (me) in understanding Haig’s concern with “chronic landslides” in which the volume of landslides at a given area increases with successive events as compared to common landslide events where the volume of landslides decreases in successive events. The crucial difference seems to be that the hierarchy is such that with successive landslides the barrier (asperities, if that is the technical word) to landslides increases in ordinary landslides (first slides involved the most stressed) while in “chronic landslides” the removal of a small stress releases a bigger stress and causes a bigger landslide.

I would think that the direction of entropy-change would be the same in all landslides; it is just the hierarchical sequence with time that changes.

I think that a stretched exponential description of landslides is consistent with the sandpile model. In the earlier blog, I have sought to make a connection (I still think it is not a wild unfounded connection) between the log-log frequency vs size plots of avalanches in earthquakes and a logs2D (s2D conductivity of a two-dimensional sheet) vs 1/Tb (T =  temperature) plots near an “insulator-metal” at which the temperature-dependence of conductivity changes sign at a critical concentration, nc, of dopant. The insulator is one in which the conductivity decreases with decreasing T or increasing 1/T.

This connection becomes reasonable once some mapping is made. The first of these is that in insulators, the electrical conductivity is a measure of the frequency of electrical charge transport across a barrier. The second of these is that in insulators the rate of transport is given by an activated term exp(-Eact/Tb) term (b = 1 in ideal insulators) where Eact is an activation barrier which is positive. The value of Eact/Tb is thus really a measure of the size of an effective barrier to charge transport. This barrier increases with decreasing T. One may relate these barriers to asperities in landslides and earthquakes. One expects that the magnitudes to the barriers or asperities determine the size of earthquakes. The logs vs 1/Tb plot may then be seen as a log-log plot of frequency vs size in the insulating phase.  

As mentioned in the previous blog, the sign of the exponent b changes sign in the so-called metallic phase. If we extrapolate the arguments above for the “insulating” phase to the “metallic” phase it would mean that Eact/T-b would effectively be reduced with decreasing T. The conductivity could therefore be expected to increase with decreasing T, as it does. The crucial factor is that the sign of the exponent has changed. One has no evidence from sandpile studies that the exponent obtained from log-log plots of frequency vs size would change sign! If it did, it would imply that the frequency of earthquakes would increase with size. This is a doomsday scenario.

One of the scenarios associated with doomsday-like disasters is when there is heavy rain. It is this that we are most unprepared for. This helplessness, is expressed, for example in learned articles (Nat. Hazards Earth Syst. Sci., 10, 2341–2354, 2010) by, say, “On the other hand, larger phenomena (e.g., rock/earth slide, earth flow) usually affect the same zone repeatedly: short-lasting phases of mobilizations are commonly triggered by external causes (like heavy rainfalls or earthquakes) and may give rise to quite limited displacements; sometimes, the area involved in a landslide reactivation results remarkably enlarged.

In the 2011 article by M. Haigh and J. S. Rawat (now Director, Centre for Excellence for Natural Resources Data Management System (NRDMS), Almora, Uttarakhand) cited above they write on finding solutions to recurring Himalayan Devastations after the 2010 Almora event. They express in their helplessness by noting that “… this study generates as many questions as answers, which is not an unusual outcome of field research”. They further find that “…these landslides were produced by hillsides that are close to their critical limits of stability and where individual sites have such a sensitive dependence on initial conditions that landslide emergence is predictable only in statistical terms …”, That is as close to their saying honestly that they don’t have a clue as to what preventive steps should be taken.

Tread Softly if you Must.

The spatial distribution of the landslides of a rock formation is controlled by its physical characteristics (lithology) such as its morphology and the nature of its water resources (hydrology). Clayey and sandy formations of the Cretaceous (150-80 milion years ago) and Paleocene (65-55 million years ago) and Eocene (55 to 40million years ago) are the most affected producing both the usual (translational) landslides and the”chronic” (usually rotational) landslides. These are roughly the periods in which the lesser Himalyas have been formed. Most slope failures takes place where there are loose rock and soil that the base of a steep slope. Some of these slides also affect the underlying clayey formations that could have formed, say, by weathering. The severe landslides occur when there is heavy rain and the clayey formations sandwiched between harder bedrock-like become muddy and soap-like. Such a situation is envisaged in the model for the formation of the Himalaygiven in Fig 6, right. In this model, steeper slopes have more loose-rock and larger slabs of  bedrock which would come sliding down altogether causing the major devastations.  This is more likely to occur in areas where the average slope is greater than 35o.

The point is that there is not much we can do about making the threshold slopes of the Himalayas get out of its precarious balancing act and instead become considerate and get into a more stable configuration for the benefit of the aam-aadmi (in its urbanized middle class sense).

It is not so much about whether it is these people require caring about. It is more about whether the government would still persist in pandering (indulging in somebody's weaknesses or questionable wishes and tastes) to their wishes.  It is more about the reckless development merely for the so-called benefit of the people in the plains and their frivolous electricity consuming pastimes. Rge trouble with road- and/or dam-buiding is that they have to blast at times. In thrshold landscapes in a self-organized critical state such blasting would sens shudders through the hill slope and loosen it furher and increase the range of criticality. Devastating lanslides or snow-slides will become stronger should it be triggered by a later heavy rainfll/earthquake event. In a Bling satellite picture (Fig 1, right) they have shown a road (NH109) tat goes beyond kedarnath towards Gangotri. I have looked closely at the maps available and find no evidence of a road being built. Thankfuly! But then, I am not sure if these maps are the most recent.

If it was the United States of America, they look after their own citizens.  Mass-scale deaths of the kind witnessed two months back would have resulted in man-slaugther litigations. The pollution in the Gulf of Mexico by British Petroleum in 2010 led to a penalty of 4.5 billion dollars (~ 25000 crores). If I take Wikipedia into account, this sum is nearly 5 times the cost of building the Tehri dam. This penalty should be sufficient to discourage developmental recklessness.  

On top of this a government such as that in USA would have filed a case of manslaughter, say, of the second degree (unlawful killing of a human being without malice either express or implied) or third degree (causing the death of another person either through criminal negligence or through the commission of an unlawful act not amounting to a felony). It can be imposed not only on the companies, but the experts and engineers who approved some dam and road-building projects. That is a very important and integral part of the democracy in USA.

If Indian legal firms are unwilling to take up this exercise because of their Himalayan-scale fragility, one may as well import legal expertise to prosecute the case. This is one foreign expertise  that I would love to have.

It is also finally dawning on me that the reason why the number of deaths in the recent Uttakrakhand devastation is being played down.

Influence of “Local” People.

There is an intriguing sentence in Haigh and Rawat’s article where they write that “…determining causes can be much more difficult, especially when the local community believes that it knows, already, what are those causes and their implications.” I don’t know whether it means that the local community resent the interference in their community in the name of development. It could be just the opposite! I have heard stories that environmentalists who oppose large-scale developmental work are harassed and chased out. The local community, if they are contractor-influenced, may actually dislike suggestions that the hills are unstable and large-scale operations should be handled carefully.

In my first Blog (29th June 2013) on the Uttarakhand Devastation titled “Colossal Himalayan Slides, Massive Pilgrims’ Stress and lessons from History: Gaia as Ardhanareshwar.” I had written in a section “Where Men are gods” the following:
It is good to remember that the Uttarkhand areas are not peopled by people from the plains. They have Tibetan (Kipling would call them Esquimaux) or Khasi fatures. The region around Nainital itself was called Khasi-desh. The Khasis still constitute nearly 40-50 percent of the population (as Khasi Brahmins or rajputs). These are people who are from the hills, who are used to the hills, are comfortable in the hills. They would not require vehicles to travel ten-twenty miles in a day. They do not require roads the way people from the plains do.

A large part of the Uttarakhand people are, on the other hand, people from the plains who went to the hills to avoid what they perceived as Islamic persecution. There were also people who went there to make a better living or because they went there many centuries ago because of the many good qualities (bahuguna) perceived in them by the rulers of the kingdoms. These people inhabit the lower stretches of the Garhwal and wield disproportionate economic influence.

It is the people from the plains who could like to be economically rich.

There was a recent (Independence Day) TV program on Channel V to raise money for the defence personnel who died in rescue attempts during the truly mind-bogglingly heroic Uttarakhand Rescue Attempts. This program was anchored by someone who used the superlatives and gestures the same way he used them for IPL matches. This is alright, I guess, if style is more important than the substance.

In this program there was a report by a TV correspondent about the tough times faced by hoteliers, transporters and restaurateurs in Rishikesh and Haridwar because of the lack of business from tourists and pilgrims. He urged, seeing the poor state of this business men, that the government (who else?) develop the roads as quickly as they can so that they can go back to their old days. The concerns are similar to that expressed in the New Yorker cartoon below where a natural disaster (balancing rock fell on the natural bridge) has affected tourist inflow.


I have no specific comment to make on this, but my scientific concerns --- for whatever it is worth and  to which I have to be committed till other evidence comes around --- tell me such a path should be completely avoided.

These businessmen are descendants mainly from people in the plains. They are different from the people from the hills, the Rawats, and the Negis and the Bhutias, for example, are used to living in the hills. They probably have more affinity to the bon people from Tibet and, from my experience, they resent being called bahadur, a term by which insensitive plainsmen call people from Nepal. These people live in a completely different style in the hills away from the pilgrims’ route. They were also among the last, I think, to be evacuated.

These hillsmen are those who realized the power of trees in holding the soil and preventing landslides (realized later to be true in a study of the Oregon ranges, and discussed above in the context of Fig 7) in their chipko movement that delayed deforestation. Except for those of the hill people that live parasitically from the tourist, pilgrim and infrastructure business, I do not find it likely that they will be unhappy with programs that take a second look at road and dam building.

I have many well-meaning friends who say that they cannot stop the clock and go backwards. Human progress cannot be stopped. It’s true in some cases. Ask the Americans and the Chinese!

A not-too-improbable Solution.

I guess we have to find our own solutions. In the context of solving the problem of the pilgrim’s progress on threshold landscapes, I think we have an easier solution, that comes from my experience of the devotees of Balaji, the Tirupathi Venkateswara. In Pune, there is a famous Kapurthol phata  on the Pune Bangalore highway (NH4) which one uses for going to Saswad or the historically important Purandhar fort and an ancient temple at Narayanpur. The Rao family of Venky’s chicken fame have their chicken feed factory at Naryanpur. Behind this factory, with the majestic Purandhar hills as a backdrop theyhave set up another Balaji mandir which is said to be a replica of that at Tirupathi.

A blog from Ruhi’s webspace on Balaji temple has this description:
“We reached Balaji and parked our vehicle in parking area. Balaji temple is replica of actual Tirupati Balaji temple in the south. The majestic walls in white marble and typical Gopur structure are amazingly beautiful. We joined the queue and after going through all those security measure we finally reachedgarbhalay. The standing idol of Balaji is royal and awesome. The upalay(surrounding area around main temple) has small temples of deities like Mahalakshmi, Padmawathi, Krishna, Kuber, etc. We took prasad and enjoyed divine and clean ambience at the temple.”

It seems to me that this example of Balaji Mandir provides a solution. One has to make real-size replicas of the various holy spots in some of the non-threshold landscapes in the Garhwal provided that they promise to keep the old pilgrim’s route strictly out of bounds. They can have the deities installed and certified by appropriate head priests, say from the Namboodiripad communities. This whole complex could be a Disneyland like complex, with periodic safe landslides that block traffic for a few minutes for food- and drink- breaks. They could have rivers full of Indo-Ganga water (appropriately chlorinated and sanitized) with mock burning ghats. There could be artificial snow ranges and ski zones. There would be high waterfalls of the Dhud-ganga kind along with small dams for water skiing. The pilgrim crowds could be ensured and periodically visited by page-three and Bollywood/cricket celebrities.

The original holy spots should be left for the genuine pilgrims. Those extra-rich and those who would like to be exclusive may use helicopters to visit these original places provided they are not using government LTA (leave travel allowance) funding. They could participate in tree-planting.

I think that we have been ignoring for too long Siva’s dance of destruction too long. The signs are all there and we should pay heed. The cartoon below is from New Yorker,



We should not think we are foreigners in our own land.