THE SHYOK FLOOD, 1929

J. P. GUNN

I. The Bursting of the Chong Kumdan Dam.

IN the first volume of this Journal Mr. Ludlow told of his attempts to reach the Chong Kumdan Dam. There is therefore no need to repeat the details of the route. This year, benefiting by his previous experience, he was able to guide our camp straight up to the glacier.*

Personally I had only the vaguest idea of what a glacier looked like at close quarters and was somewhat disappointed to find it resembling nothing so much as a great heap of shingle with a few white points projecting from its surface. I had expected to see a large mass of white ice, with transparent pinnacles, and I was quite prepared to find the whole mass of the same transparency as ice on a pond. A little reflection showed that from the very nature of a glacier's movement, the only possible structure for its main body, was a granular form of ice, and as it dawned on me that this was a most excellent structure for localizing the effects of an explosion, I wondered whether the exponents of the " blowing-up " method of dealing with the dam had devoted much attention to the nature of glacier ice.

The dam was a mighty mass of ice, rather too big for one to appreciate its full size, but dwarfed by the surrounding hills, rising from three to six thousand feet above the highest pinnacles. Seen from the hillside above, the ice in the river-bed appeared roughly square in plan ; measurement showed it to be four thousand feet long across the bed, and about five hundred feet high.

From its size I was of the opinion that the only way failure could occur was by the water in the lake overtopping the dam ; and from the structure of the ice, I was sure that the main body of the dam would offer no great resistance when this occurred, and that a big flood was bound to be the result.

From Ludlow's information it was obvious that the lake was much larger than in August the year before. We therefore determined to ascertain its volume. After inspecting the dam from the north I concluded that it would not be overtopped for at least another two months, and that probably it would not break till the following year, since before long the rate of melting of the glaciers feeding the lake would decrease. All the time we were down at the dam on 12th August loud creaks and " groans " were heard, lasting some time, as ice-floes broke off the main body. Both Ludlow and I considered that when the dam did fail, it would break up rapidly.

When we had finished our work at the lake, Ludlow went on to Kashgar and I returned from Daulat-Beg-oldi. On my return journey, as I approached the Shyok from the ravine leading to Murgo, I noticed a few blocks of ice lying at the foot of a side nullah and looked up to see where they could have fallen from. While I was still looking up I turned the last bend and on glancing towards the Shyok, the first thing that struck me was that the shingle bar at the mouth of the ravine had changed its shape. Then I observed that the whole bed of the Shyok was covered with ice-blocks, and realized at once that the dam had burst.

The Shyok river was still in moderate flood, though this was the 17th August and more than forty-eight hours after the dam burst, as we discovered afterwards. I therefore gave orders for the zak, the local form of raft made of inflated skins, to be got ready, while I set off up the left bank to see whether there was a practicable road. The skins had not been used for some time and had to be soaked. It was therefore not till the early morning of the 19th that I was able to ferry a light camp across the river and pitch my tent at sunset some three miles below the dam.

All the way up from the river there were evidences of the magnitude of the flood. Ice-blocks up to twenty feet cube were scattered about the banks of the river, sometimes seventy feet and more above the river-level, while occasionally there was a gigantic block about fifty feet cube by way of variety. All the shingle mounds in front of the Aktash glacier had been washed away, and the terminal pinnacles eroded. In backwaters the ice was piled up in regular beaches in steps anything up to six feet high and fifty feet wide, one above the other, and from the general appearance of the river-bed, I was sure that very little of the glacier was left.

Almost every time I had guessed anything about this glacier, I had guessed wrong ; and this was no exception. From my camp I could not see the whole of the glacier, but except for a few more white patches than were visible before, I could observe no change in the glacier at all! At last some idea of the immensity of the glacier began to impress itself upon me. The channel, through which the whole of the water in the lake had escaped, was barely noticeable. Yet the water in the lake, before the flood, had been sufficient to cover an area of ten square miles to a depth of a hundred and seventy feet. I find it even now extremely difficult to convey any idea of the magnitude of the dam, or of the volume of water which passed through it ; and I must leave the photographs to show how very little the dam was affected.

On the 20th, when I approached the dam I was able to see clearly what had occurred. The dam had burst along a curved line extending from near the right bank of the lake on the northern side, through the highest portion of the dam, nearly to the left bank of the river at the southern side. The cut, through which the water had escaped, was about four hundred feet wide, and the ice stood almost vertically on either side. A small quantity of water had blown a hole in the right side of the cut, as can be seen in the left foreground of the illustration. The lowest part of the dam, the low gully along the cliffs at the snout, which I had considered the danger point, was unaffected.

In only one particular were my anticipations of the bursting of this dam correct. I predicted that when the dam broke, there would be a big flood, in all probability larger than that of 1926. And it was so.

II. The Shyok Flood in the Gilgit Agency.
H. J. TODD.

Since despatching my letter giving details of the floods which have occurred in the Grilgit Agency,16 the long expected " Shyok Flood " has come and gone, and curiously enough I happened to be the first person within easy reach of telegraphic communication to witness its arrival in the Indus, where the latter passes through the Grilgit Agency.

I was returning from Kashmir to Grilgit, and on the morning of 17th August, had left Bunji at four o'clock to ride the seventeen-mile march to Pari. On this march the road crosses the Indus river, by the Partab Pul, some seven miles beyond Bunji, and here I arrived about half-past five.

It was apparent that something was wrong. At that time of the year the Indus is about at its maximum summer flood-level, but the Partab Pul can always show a clearance of some 45 or 50 feet. When I arrived, there seemed to be little more than a 20-foot clearance, and the water was coming down in a dark chocolate-coloured flood, carrying large quantities of drift-wood, as if the river had just succeeded in washing its banks of the deposits of previous floods. Shortly afterwards, however, appeared an occasional large uprooted tree, and innumerable poplar poles, cut in regular lengths as if for building purposes. Later, boards and roofing material began to arrive, clearly indicating the fate of some unfortunate village in Baltistan.

A road gangman, who had been sleeping at the bridge-head, informed me that he had first noticed an unusual rise some three hours before my arrival.

My thoughts, of course, immediately flew to the Shyok dam. With the aid of a stone, on a piece of string, I was able to calculate that, at 5-30 A.M., the river was 18 feet below the lowest cross beams of the roadway structure.

An hour later the river had risen to within 12 feet and all doubts as to the origin of the flood immediately disappeared. A levy was sent galloping in to Bunji to wire the news to Kashmir, and Bunji and Chilas were warned to take all precautions, and to commence recording the rise of the river for the information of the authorities down in India.

At 7-30 A.M., the flood had risen another 3 feet, and at 8-15 A.M. it was only some 6 feet below the bridge.

The flood was now beating strongly against both abutments of the suspension bridge, and occasionally a high wave in midstream would dash against the bridge roadway and cause it to swing perilously.

At 8-45 A.M., the flood was within 5 feet of the bridge, but at 9-30 A.M. it was a great relief to find that there was no further rise to record, and, if anything, the midstream waves were striking the bridge less frequently.

The worst was over and the bridge was safe.

At 9-45 A.M., the levy returned from Bunji and was able to recross the bridge with his pony. He reported that the Post Commandant at Bunji had now got going, and the readings at the regular gauge there were being recorded and wired to Srinagar.

At 10 A.M., the river was distinctly quieter, and at 10-30 A.M. I had to continue my march, leaving a watch at the bridge. By 2 p.m. the watch reported a fall of 4 feet.

I always make it a practice to carry a camera with me on trek, but this morning of all mornings, my bearer had noticed that the sling was becoming unstitched so had packed the camera in my yakdan and sent it ahead ! I thus lost the unique opportunity.

The Partab bridge is the only means of crossing the Indus during the summer months, and it is of course during these months that the supplies for the Gilgit garrison are sent up from Kashmir. Had the bridge gone, a most awkward situation would have arisen, as there would have been no chance of re-erecting the bridge until the autumn at least, when it is again possible for a ferry to ply safely.

At Partab Pul the Indus flows between very high cliffs and the actual length of the bridge is 330 feet.

For some days before the flood, a phenomenal heat wave was being experienced in the Gilgit Agency, and doubtless its effect was felt at the Shyok dam. At Gilgit itself a temperature of 112°F. was recorded on 13th August, two days before the dam burst, the previous summer record being just over 105°F. For this reason the Gilgit, Hunza and Indus rivers were undoubtedly somewhat higher than usual, but the increase over normal summer level could not have been more than & few feet.

On 7th May the Partab bridge was 57 feet above the river-level and on 1st August the clearance was 51 feet. By noon on 19th August the river had returned to its pre-flood normal level, giving a fifty-foot clearance.

At Bunji, where the river has more shelving banks, and is much broader, hourly observations were taken from 7 A.M. until 10 p.m. on the 17th, and again from 5-30 A.M. to 3 p.m. on the 18th, when the river had returned to 4 feet above normal. The highest point, 35 feet above normal, was reached at 10-50 A.M. on the 17th, at which it remained for nearly 2 ½ hours. The river became normal again at Bunji at noon on the 19th.*

At Chiles the river was first reported to be rising at approximately noon on 17th August, and reached a height of 53 feet above normal about midnight. On the morning of the 18th the river was subsiding and reached its normal on the evening of the 19th. No detailed measurements were recorded and the above information is therefore only approximate.

No loss of life was reported in this Agency, and the damage to property was very small.

 

III. The Shyok Flood : A Commentary.
Major KENNETH MASON.

The bursting of the Shyok dam gives rise to many speculations : the inhabitant of the plains of India south of Attock, the hydraulic engineer and the glaciologist are all concerned. The first must have thanked his gods that the dam burst on the morning of the 15th August, and not on that of the 25th. Had it collapsed on the latter date the liberated waters would have arrived at Attock at the same time as the higher flood caused by an unprecedented rainfall in the Kabul river basin, and the combined floods must have done an appalling amount of damage in the plains. The plainsman may speculate on the likelihood of other such visitations, and he may not know that, coming as it did, some ten days before the rain-flood, the Shyok burst was a blessing in disguise, for it caused little damage, scoured out the bed of the Indus and so gave a better " run-off " for the rain-flood.

In a most interesting paper published by the Punjab Government, Mr. J. P. Gunn has investigated very thoroughly most of the problems connected with the hydraulic aspect of the Shyok flood.17 Reports and letters have also reached me from Khan Sahib Afraz Gul Khan, of the Survey of India, who was camped at Daulat-Beg-oldi, when the dam burst, from Mr. H. J. Todd, the Political Agent at Gilgit, who was actually crossing the Partab Pul, when the flood passed that place, from Lieut. J. Barron, in charge of the Kashmir State Artillery, stationed at Bunji, and from Major G. Y. B. Gillan, the First Assistant to the Resident in Kashmir. All these gentlemen are members of the Himalayan Club. I am also indebted to Mr. S. Walker, Chief Engineer and Secretary for Irrigation, North-Western Frontier Province, to Major C. F. Carson, R.E., Bridge Engineer of the North-Western Railway, and to Brigadier E. de L. Young, Chief Engineer, Northern Command, for details above and below Attock. It is not possible to publish all these communications in detail.

Khan Sahib Afraz Gul Khan at a distance of 19 miles from the dam heard the first breaking of the ice " like the noise of a cannon- shot " at 5 A.M. on the morning of the 15th August ; and during the day occasional "booms" were heard by him and members of the Yisser expedition, which they attributed to breaking ice. This points to the conclusion that the dam did not collapse instantaneously and completely, a conclusion supported by Gunn's observations at the dam itself. The form of the " wave-curve " at Saser Brangsai, tau miles below the dam, shown in the accompanying chart, indicates, however, that the first break mast have been the main one. If the records gleaned by Gunn from the ford-guard at Saser Brangs;a ere trustworthy, it appears that a channel was probably cut through the top of the dam at first, and that a second break occurred, probably about 11 A.M.18 This second break allowed the lake to be completely emptied, and the waters released by it seem to have arrived at Saser Brangsa about 4-30 p.m.

Gunn, who had been sent by the Punjab Government to investigate the Shyok dam, and who was accompanied by M[r. F- Ludlow, whose account of the situation in 1928 was published in the first volume of the Himalayan Journal, examined the dam from the lake side less than a week before it burst. He records that the lake at the time of release was approximately four hundred feet deep the dam, and that its volume was approximately 1,095,500 foot-acres. The main channel through which the water escaped was surveyed by him afterwards and found to be only four hundred feet wide and to have vertical sides. This channel was cut diagonally from near the right bank of the river upstream of the dam, and emerged on the downstream side near the left bank. There was no break adjacent to the wall of rock against which the ice impinged, and where weakness seems to have been expected owing to radiation. The incomplete burst at first and the narrowness of the channel must, I think, account for the time taken for the flood to cover the first ten miles to Saser Brangsa, where the rise was only 85 feet in four hours.

From this point the flood was contained in a valley with a, steep fall, and when it reached Khalsar, 135 miles from the dam, it had concentrated into one huge turbulent wave, which passed that; place almost entirely in eight hours, the maximum height of 63 feet being reached in two hours. Gunn surmises that this concentration w^as assisted by a large shingle bar, which was removed during the passage of the flood, and he records that the bed of the valley at Khalsar scoured out to a depth of nine and a half feet. He also gives some most instructive comparisons between the 1926 and 1929 flood-marks, between Khalsar and Deskit.

Gunn remarks that lie and others (amongst whom I must plead guilty) over-estimated the reservoir effect of the broad Nubra valley, which enters the Shyok a short distance below Khalsar. The 1926 flood, he records, did not reach the village of Burma, some five miles up the Nubra, and on the right bank ; while the 1929 flood destroyed the village. The reservoir formed in the Nubra by the flood of the Shyok was a mere five thousand acres in extent. This reservoir question is one for the hydraulic expert; but whether the Nubra valley above the Biagdangdo gorge, or the open ground near Khapalu below it, was responsible, the fact remains that the rise at Skardu, 175 miles below Khalsar, was only 25 feet, in spite of the fact that this is below the confluence of the Indus near Gol. The main flood took eighteen hours to pass Skardu, as against eight hours at Khalsar.

To my mind, the extraordinary feature of this flood, which I, and I believe others, failed to realize beforehand, was its amazing recuperative power between Skardu and Partab Pul. These places are 137 miles apart, and they are separated by a very confined V- shaped water-worn valley. While the height of the flood dropped from 63 feet at Khalsar to 25 feet at Skardu, after the passage of 175 miles, nevertheless it rose again to 45 feet at Partab Pul, 137 miles below Skardu. Partab Pul is just below the great " knee-bend " of the Indus, at which point the flood seems to have banked up considerably. Mr. H. J. Todd's account of its passage between the cliffs and under the bridge here gives some idea of its turbulence. Though he records that prior to the flood, the Indus, Gilgit and Hunza rivers probably were a little higher than normal, owing to the phenomenal heat-wave, this increase can have had very little effect in raising the flood-level.19

In the broad Bunji valley the maximum height of the flood dropped rapidly to 35 feet in seven miles, as indicated by the very full series of observations taken hourly at Bunji itself. But if reliance can be placed on the scanty records at Chilas, it had, after the passage of another restricted valley, again risen to 53 feet, only 45 miles lower down. Below here the Indus enters the most confined gorge of its whole course, and we can only conjecture what height the flood may have attained. At Torbela, about 696 miles from the dam and 243 below Bunji, where the Indus again flows through more open country, the rise was only 16 feet ; at Ghazi, 13 miles lower, still 14.5 feet ; and at the Attock bridge, 742 miles from the dam, where the river is temporarily confined again, 26.9 feet. Gunn calculates that of the total volume of 1,095,500 foot-acres of water impounded by the dam, 884,300 foot-acres passed Attock in three days. Below Attock the flood soon was dissipated; at Bilot, 160 miles further, its height was only 2-76 feet, while the rise at Dera Ismail Khan was only 1.6 feet.

I have gone into these figures at some length, because those at Partab Pul, Bunji and Chilas do not seem to have been available when Gunn wrote his report. They emphasize in an extraordinary manner the recuperative effect of floods when passing through gorges, especially when they are markedly V-shaped, and when they have no reservoir, such as the Nubra valley above them. They show also how rapidly the height of the flood falls when the waters flow between shelving banks. The upper currents during the passage of a gorge must be far swifter and more violent than those below, to enable the rear portion of the wave to superimpose itself upon the peak. I doubt whether anyone anticipated that the height of the flood-wave at Attock would actually be greater than at Skardu, 432 miles upstream of it !

 

I calculate the velocities of the flood peaks between the different stations as follows :

Saser Brangsa, approx. alt. 15,200 ft.

Gunn notes that the damage done by the flood was not excessive, considering the huge mass of water liberated when the dam burst. He records that most of the damage was done between Tirit and Skardu, that the total cost of it in the Skardu tahsil was estimated at under three lakhs of rupees (£22,500), and that on the whole river only forty-eight villages were affected. As far as I have been able to ascertain, there was only one life lost; the son of the ferryman at Bunji was swept away and drowned while trying to retrieve driftwood from the flood.

In his report Gunn discusses the probability of the dam reforming and impounding another lake. This question is more in a glaciologist's sphere than in that of a hydraulic engineer. It is in fact of absorbing interest to the former. Gunn adduces, as evidence of his conclusion that the Chong Kumdan glacier is probably now in retreat, the behaviour of the adjacent glaciers, the Kichik Kumdan and the Aktash. These two glacier-snouts are known to have been recently retreating. It is, however, a fact that of two neighbouring glaciers, even when they have the same exposure, the snout of one may be advancing while that of the other retreats ; and there are even instances of one side of a glacier-snout advancing while the other retreats. The matter is decided mainly by the time taken for an excess or defect of snowfall above the snow-line to reach the snout, and with compound glaciers particularly, the conditions affecting the flow may be most complicated. In a recent investigation that I have made into the snout variation of thirty-four glaciers of the Karakoram region,20 I have tried to explain the four components of snout-movement, viz., secular, periodic, seasonal and accidental, and I have given the various conditions which I believe enable one of these components to preponderate over the others.

Nor can the absence of detached pinnacles be regarded as evidence of advance when the snout is eroded by a powerful river. Creaks and " groans " near the snout either in winter or more particularly in summer are generally signs of advance, but here too when the glacier is holding up a great mass of water, these noises may be caused by the pressure of the impounded lake. A flattened blackened snout in summer or an upstanding white snout in winter are normal seasonal signs and cannot be brought as evidence. The absence of a terminal moraine is easily explained by the erosive power of a river issuing from or passing the snout.

The solution of the problem of whether the Chong Kumdan glacier will impound another lake is therefore not easy. It is, in my opinion, no use examining the adjacent glaciers ; it is a personal matter which concerns the Chong Kumdan itself.

The fact that the pressure of the glacier from the valley behind has been able to keep closed all crevasses in the portion of the glacier stretched across the Shyok valley is an argument that at least until recently the glacier has been trying to advance. The fact that when Gunn first examined the dam from the south side he saw neither an issuing stream due to percolation through the glacier, nor a terminal moraine, which must have been visible if there was no stream to carry it away, and if the glacier was in retreat ; the very fact that that the dam did not break at the point of junction with the marble rocks, the radiation from which must tend to separate the ice from them at the first release of pressure ; and the appearance of the ice in his photographs prior to the burst; these three factors all denote a tendency to advance. The dam did not burst until the level of the lake nearly reached the summit of the dam; if the pressure on the opposite wall of the valley had been relaxed by retreat, I would have expected it to have burst before that height was reached. I do not therefore think that the arguments that Gunn puts forward are sufficient to conclude definitely that the glacier has begun to retreat.

The Chong Kumdan glacier is one where the periodic component is markedly predominant; and it is fortunately one, as I explained in the last Himalayan Journal, where we have observations and deductions extending over a whole century. These observations indicate a periodicity, I believe, of approximately forty-five years. If the snout were not eroded by the Shyok river nor held up by the cliffs on the left bank of it, its years of maximum advance would be approximately 1839, 1884, and 1929. Examining the records of these years and the years near to them, we find floods due to the bursting of the Chong Kumdan dam occurred in the following years : 1835, 1839, 1842, 1926, 1929. About the year 1884 we have no actual records of floods; but we know that in 1873 Colonel Gordon of the Forsyth Mission to Kashgar recorded that the Chong Kumdan " almost touched the opposite side of the valley," and that for the next fifteen years we have no records of any traveller using the valley route. On enquiry from the Chief Engineer of the North-West Frontier Province, I learn that an abnormally high flood occurred at Attock between the 11th and 20th August, 1879, and another record flood occurred at that place on the 29th July, 1882. These two floods were higher than any others for several years before or after, and the second was actually higher than the Shyok flood at Attock this year by five feet. I feel convinced that these two floods were caused by the bursting of the Shyok dam during the glacier's last period of advance.

From a study of the periodicity curve of the Chong Kumdan glacier and from the other factors noted above, I therefore come to the conclusion that the glacier is at its maximum periodic advance. It must, however, be remembered that in August surface ablation is at a maximum and therefore seasonal degeneration and seasonal retreat have weakened the glacier and magnified its weak spots. There was an abnormal heat wave for a few days prior to the burst, In my opinion the normal seasonal advance and regeneration in the coming winter will almost certainly close the narrow transverse channel that has been cut, and by next spring this should have completely healed. I believe that another lake will almost certainly form next spring, but since the seasonal retreat next summer will now be assisted by periodic retreat, the dam will definitely degenerate in height and strength. It may be that the lake so formed will drain away by percolation, or it may gradually wear away a channel, taking several days to drain (as happened in the last of a similar series with the khurdopin glacier). If the healing is so complete as to prevent either of these two courses, I believe the dam will break again in August 1931, the month of maximum inflow to the lake, and the month of maximum degeneration of the ice. Under no circumstances can the dam impound a lake in the next few years of the same magnitude as that liberated in 1929. In my opinion there is no danger of a serious flood for many years to come, while the present danger of a complete block and of a minor flood will be over at latest in 1932, after which there will be nothing to worry about till 1969. There will then be a eight-year scare-period.

In Figure II, I have shown the periodicity curves for both the kumdan glaciers, which, I hope, will illustrate my argument. The observations of the snout of the Kichik Kumdan are much less reliable than those of the Chong Kumdan, and its snout is much more subject to variable end-erosion by the river, and to accidental retreat (but not advance), owing to idiosyncrasies of the Chong Kumdan. I believe, however, that its period is also approximately forty-five years and that this period is entirely out of time with that of the Chong Kumdan, one being at its maximum advance when the other is approximately —at its maximum retreat. There is only one recorded flood from the Kichik Kumdan, and it does not appear to have done much damage, this is accounted for by the fact that (to use Gunn's words) " the configuration of the country is such that even if it did impound a four-hundred foot depth of water, the volume of the lake would only be about one-third of the volume retained by the same depth at the Chong Kumdan."

 

Postscript.

Since going to press, Gunn, who is in England on leave, has kindly sent me the following comment on my paper : " I do not appear to have made myself quite clear in the report ; the water never got within fifty to a hundred feet of the top of the dam at the lowest point, and was probably a hundred and fifty feet from the top at the point where it burst. The gradual increase of discharge was undoubtedly due to the opening increasing in some way—probably widening—and the sides falling in would account for the further noises heard by the Vissers. At the same time, on the 10th I heard a tremendous roar, but there was nothing visible at the dam to account for this when we saw it on the 12th. The size of the channel is only an estimate, and it was not actually surveyed.

“The variation in size of the flood rise is quite natural and depends on the available waterway, being much greater in gorges. The rise at Skardu is affected by the water being able to back up the Indus at the Indus-Shyok junction, which reduces the height but increases the duration of the flood.

" So far as I gathered, the total loss of life was one man at Deskit, who tried to rescue a pony, and eight women—three in one village and five in another—somewhere about Abdin, above Skardu.”

The above statement that the dam burst before the water-level of the lake reached within fifty to a hundred feet of the top of the dam at the lowest point indicates, as Gunn suggests, that the snout has probably passed its position of maximum periodic advance. Periodic retreat however sets in very slowly and there is every chance of the winter advance counteracting the first tendency to retreat. It must also be remembered that the upper layers of glacier-ice are less compacted and therefore weaker than the lower ones, so that the break may have occurred along a line of weakness.

⇑ Top