Transactions and Proceedings of the New Zealand Institute/Volume 5/Article 10
Art. X.—On the Taieri Floods. By G. M. Barr.
[Read before the Otago Institute, 12th March, 1872.]
Many schemes have been suggested for the prevention of the floods which have been so disastrous to property on the Taieri plain, but those which were generally recognized as the most practicable have usually been estimated at so great a cost as to exclude the probability of their being carried out either by the residents on the flooded area or the Government; while there is one which has been mooted upon several occasions, but dismissed almost as soon as mentioned, which appears to me worthy of more consideration. Having had for some time the feeling, scarcely raised to the position of an opinion, that the proposal to form a store reservoir at the Taieri Lake had not received sufficient attention, I took the opportunity, while in that neighbourhood last December, of spending some hours in making a careful examination of the physical features at the outfall, and I now propose to investigate how far the damming back of the waters of that part would tend towards the prevention of those disastrous floods to which the Taieri plain has been subjected in late years. Unfortunately, I have not at command sufficient information either as to the rainfall or the configuration of the ground, to give exact quantities in dealing with the whole of this subject; but for purposes of preliminary enquiry we may find enough either from direct observation or from general laws which may bear upon the subject. In the following calculations I have been much indebted to the elaborate survey executed in connection with Mr. J. T. Thomson's report upon the subject in 1870, which has been kindly placed at my disposal. Other parts are filled in from the general map of the Province, and by personal observation.
Before considering the case I shall glance briefly at the nature and proportions of the evil, for without a knowledge of these we cannot judge of the feasibility of any proposed remedy.
Referring then to the map of Otago, we find that the lower Taieri plain lies at the mouths of the Taieri and Waipori rivers and the Silverstream, all discharging large quantities of water during floods, especially the former river, which has an outpour per minute through the gorge at Outram, even at its lowest, nearly equal to that of the Clyde in Scotland between Glasgow and Port Glasgow; but having a fall very much greater, the cross section is correspondingly less. By computing the drainage areas of these local rivers, we find that into the basin occupied by the Waipori and Waihola lakes, and the Taieri plain, there is discharged that portion of the rainfall over 2065 square miles of country, which has escaped evaporation or absorption by plants or porous strata, the relative areas being—
| Square Miles. | |
| Taieri river | 1,730 |
| Waipori | 265 |
| Silverstream | 70 |
| 2,065 |
The Silverstream being comparatively small, and the waters of the Waipori being discharged into the lake of that name, with an effect upon the floods only in connection with the Taieri, I shall simply refer to them at that stage when we come to view the storage room on the lower parts of the plain, and shall devote this portion principally to the nature of the larger river and its floods.
A comparison of the Taieri with the chief rivers of Europe or North America will show that in proportion to the country drained it is much in excess of the most of these as a flood producer. Thus, while it pours down its water at the rate of 1.666 cubic yards in the second for each square mile drained above Outram, the Mississippi in flood flows only at the rate of 0.044; the Ohio, at its mouth, 0.122; and the Yazoo 0.372 cubic yards per square mile drained. In Great Britain, the Tyne is the only one which approximates to the Taieri in this respect, its rate of flood discharge being 1.12; but small streams which have been gauged in meadow land have yielded as much as 1.2. The Yellow River, in Ireland, is as high as 4.12. On the continent of Europe the Loire appears to be pre-eminent for its floods—yielding at Pont de Fleurs as much as 4.18 cubic yards per second for each mile drained. Many of the Indian rivers far exceed any of the above—the Irvitz, especially, delivering at the rate of 16.5 cubic yards per second for each square mile of its gathering ground.
Now these figures of course represent very rough comparisons between the several streams named, no two of which are alike in physical conditions, either as to amount of rainfall or configuration of drainage area; but in a general way they enable a classification to be made which may yet be further improved as information is gathered. Apart from the relative quantities of rainfall, there are many other circumstances which tend either to aid or impede floods. Thus a stream draining a large tract of country is much less liable to heavy floods than one draining a small one; and also the general nature of the country as to inclination has a most noticeable effect upon the rate at which the water finds its way to the river, and consequently a like effect upon the amount of its volume. Another most important consideration is the nature of the strata, or the amount of vegetation in the district drained; for when the rocks are of a close compact nature, with comparatively few joints or crevices, and the ground bears but little vegetation, the water will run quickly oflf; but where the ground is porous, and the vegetation rank, a much longer time will elapse before the underground basins are filled, and the ground so saturated as to shed the water off as the rain falls. Surface lakes have a comparatively greater effect in moderating floods than either of the above causes, by receiving the water as it comes from the creeks, and allowing it to spread in thin films instead of rushing down a river channel in deep volumes.
Having got these general results, we may now endeavour to examine particularly the conditions of the Taieri drainage ground, with a view to ascertain, if possible, in what respects it facilitates the rapid discharge of the rain or melting snow, and it will then be seen that the chief features likely to promote this are the generally mountainous character of the catchment area, with the exception of the Upper Taieri plain, the steep inclination of the ridges, the nature of the rocks, and the general steep declivity of the bed between the upper and lower plains. On the other hand the upper of these plains forms a natural basin about 280 square miles in extent, but of this only about one and a half square miles are an open lake, the remainder being deposits of shingle of various depths lying upon impervious clay, and capable of being a store reservoir only to the extent of the interstices between the stones. That portion, in fact, resembles a huge sponge, acting with the open lake in retaining the water, and preventing to a certain extent its sudden rush down the channel towards the lower parts. The lake lies at the flank of the Lammerlaw or Rock and Pillar range; and finds along with the Kyeburn an outlet through a narrow gorge, at one place not wider than 110 feet, but unfortunately its low level limits the capacity of the whole reservoir as a flood moderator. The point now is, to consider the practicability of raising that outlet to such an extent as to store the greatest flood waters that are likely to occur, and release them only at such a rate as shall not be prejudicial to the low grounds lying below Outram.
The flood of 4th February, 1868, being the greatest on record, I shall take as the standard one, seeing that it is necessary in any remedial works that may be proposed to provide against the occurrence of an evil at least equal to that already experienced.
The circumstances of rainfall attending the floods of January and February of that year, at least so far as we can judge the Taieri basin by observations taken at Dunedin, were almost such as to lead us to expect that no such floods are likely to occur for very many years. Upon examination of the meteorological tables prepared by Dr. Burns and the Meteorological Department, it will be found that the unprecedented nature of the January and February floods was more owing to the extreme degree of saturation in which the ground must have been by months of previous rain, than to even the heavy rains of any one particular day. Doubtless these were very heavy, but not so much so as several days both before and since, which did not produce the same rise in the water. For the months of October, November, December, January and February, the rainfall ranged from 5.0 to 8.078 in., thus showing a continuance of wet weather unequalled, I believe, in the history of the settlement. In the case of the flood of January, 1870, which did not rise so high as the one we are specially considering, the rainfall recorded in one day exceeded that shown upon 4th February, 1868, but that followed some months of comparatively dry weather, though it commenced one showing a total of 7.399 in.
Floods may also suddenly arise after a drought, from the fact of some kinds of soil being so thoroughly baked as to have a surface almost as impervious to rain as rock itself, and thus it delivers the water almost as it falls; but such floods will be of short duration, and only until the surface has had time to soften a little and allow the usual amount of soakage.
I have referred to the rainfall at Dunedin, because unfortunately we have no records of its amount for the Taieri basin, and if we were wanting to go minutely into the matter, as we shall see immediately, the rain gauge in this city would be apt to mislead rather than enable us to arrive at correct results. Those who have paid any attention to meteorology must be aware of the great diversity of rainfall in different districts, even within a few miles of each other, and consequently of the necessity for separate observations in various localities, if any practical result of value is to be evolved. We are safe enough, however, in taking the evidence of the Dunedin observations to prove generally a season of unexampled moisture within the Taieri basin; but it will be seen by a study of the figures representing the flow of the Taieri during the flood of 1868, that for particular occasions they are of little value. Thus, for the flood of January 28th, a fall of 1.648 in. was shewn, yet that did not raise the Taieri river so much as a fall registered as 1.37 in. did a few days afterwards. This, however, does not completely prove the position that the Dunedin register is different from what the Taieri one would be, for it is still possible that the whole area of the gathering ground was not saturated to its fullest extent on the 28th January, so that a greater degree of saturation, as on 4th February, would greatly aid the rain of the latter date in producing a flood greater than the one a few days previous. Better proof, however, is found in referring to the delivery by the river at Outram, which, as recorded by Mr. Thomson, was at the rate of 4,653,068 cubic feet per minute; which would show a fall at the rate of 1.67 in., instead of 1.37, even with the whole water run off to the river just as it fell. Even this, however, is not quite satisfactory, for it is quite possible, and indeed likely, that in both localities there was a space of time—perhaps extending to hours—in which the rainfall was of greater intensity than even the higher of these sums represents. As it is impossible, however, to have the records for each few hours, we must generalize from the most frequent, viz., those for each twenty-four hours.
I shall now endeavour to ascertain the length of time which may be considered as the duration of the flood, or how many hours elapsed between the moment when the banks overflowed, and the outlet to the sea was too small to allow the whole waters which were issuing upon the plain, and the Waipori and Waihola lakes, to pass off.
For this purpose it will be necessary, in the absence of particular data, to assume that the Waipori river and Silverstream delivered their waters at the same rate as the Taieri, in proportion to area; and for the general purpose of this investigation that will be enough. This would show a total outpour from these sources of 5,532,900 cubic feet per minute; and deducting from that amount 1,186,900 cubic feet, which could flow towards the sea, the amount of water which would be dammed back would be 4,548,000 cubic feet per minute, representing the rate of rise of the flood. But it has been ascertained that on this occasion 4,585,996,800 cubic feet were the total flood waters, so that dividing the one quantity by the other we would have a period of flood equal to seventeen and a half hours. Of course, this is to be regarded as simply a hypothetical statement which will represent only the average rise and length of time resulting from that, for it is very likely that the rise would be at this rate only for a few hours, which would consequently necessitate a more lengthened period to produce the same total accumulation from a smaller rate of increase.
We have now to consider the amount of water which it would be desirable to prevent flowing upon the plain by the Taieri river, in order that the floods may not attain to such an extent as to overflow the banks. It has been found that the damage caused about the West Taieri has been the consequence of the small section of the river further down not being sufficient to carry off the whole waters as they arrived; and it has been shown by Mr. Thomson that the smallest section has been able to pass down 1,173,744 cubic feet per minute; but it might not be safe to charge it with even this quantity, so we will leave a considerable margin by taking it only to the extent of 900,000 cubic feet, looking to store the remainder in the Upper Taieri lake, or upon some of the tributaries of the river, such as the Deep, the Sutton, and the Lee streams. It will be observed that this makes no provision for the waters of the Silverstream, or the Waipori river. The first of these could probably be stored in the lagoon near its junction with the main river, and which is evidently the natural flood moderator of that stream, or in some reservoir higher up; while the large flow from the Waipori would be allowed to spread over that and the Waihola lakes. The general result, then, may be arrived at thus:—
| At Feet per Min. | |
| The quantity poured down the Taieri | 900,000 |
| „„„Waipori | 730,000 |
| 1,630,000 | |
| Outflow to sea | 1,187,000 |
| 443,000 |
So that the flood, under these conditions, would accumulate at the rate of 443,000 cubic feet per minute, which would spread over the areas of these two lakes, and raise their surface about three feet nine inches in a twenty-four hours flood. Even though there were such a rise, no serious damage would result to the adjoining lands; but in this calculation there are two elements, which are taken at extreme figures, viz., the length of time, and the rate of discharge by the Waipori, which is much in excess of that estimated by Mr. Thomson, so that we may reasonably presume that even under such circumstances as those of the 1868 floods, the rise would be very much less than that I have stated.
The quantity to be stored above Outram would evidently be the difference between the largest flood delivery there in the 1868 floods, and the quantity which I have already named as likely to get past the least capable section of the river, between there and the East Taieri bridge, amounting to 3,200,000 cubic feet per minute. Now this would be derived from different districts, the comparative areas of which are—
| Square Miles. | |
| Above Taieri lake and Kyeburn | 850 |
| Sutton, Deep, and Lee streams | 480 |
| East of river between Lake and Outram | 370 |
| 1,700 |
But the configuration of the country to the east of the river is unfavourable for storing a large quantity of water upon any of the tributary gullies, and the basin of the Taieri lake is more favourable, so that we must calculate upon having none upon the eastern portion, but store a correspondingly larger amount upon the Taieri lake. Taking therefore three-fourths of the total for that part, and the remaining fourth to be stored in small reservoirs upon the Sutton, Deep, and Lee streams, the Taieri lake would require to have its outlet so raised as to enable it to store as much as 3,002,400,000 cubic feet in eighteen hours. Now the area of that lake and part of the Kyeburn valley which would be affected so far up as the present ford on the Dunedin road is about 91,846,260 square feet, so that the increase of depth would amount to thirty-four feet. This additional depth would be required over the whole area named; but owing to the fall of the valleys, this could not be got on an average without raising the dam to a height above the bank at the bridge, and thus requiring a considerable extension in length. Probably a more economical method might be by a smaller dam at that part, and other two at the outlets from the lake proper, by which means suflicient storage might possibly be obtained for the waters of floods less than that of February, 1868, and also sufficient to reduce a similar one to safe limits, for if they could be kept back for even twelve hours, their most destructive effects would be moderated. It would also be necessary to have reservoirs upon the Sutton, Deep, and Lee streams, perhaps more so than in the Taieri, in proportion to their areas, as the features of their catchment basins are such as to show many indications of rapid flood-producing streams.
The mode of flood prevention I have examined in this paper is one which has been much adopted upon the continent of Europe, and notably upon the river Loire, which I have already referred to as standing remarkably high as a flood-producer. Above the particular part where the discharge I have referred to was gauged, we have seen that it ranks nearly three times as intense as the Taieri; yet to moderate these waters a weir sixty-five feet high was erected in 1711, which did immense service in the floods of 1846. They topped it, however, by a height of about five feet, but were still sufficiently restrained to lessen considerably the damage which otherwise would have been sustained.
The advantages which the prevention of the flow of the waters upon the lower plain possesses over any scheme of embankment, either along the present channel or any new one, are so evident as scarcely to require remark. Besides being much cheaper, it possesses an advantage in this, that even if carried out to a partial extent it produces general benefit to all the land which has hitherto been liable to inundation; but by the method of embankment upon the plain, intended to shut the water off particular parts, these portions are protected only by aggravating the evil upon other spots, both by the increased depth of the water and the heightened current.
One objection to this method has been so often urged that, paradoxical though it may appear, I believe that had it been founded upon facts, they would, ere this, have been recognized as an argument for its immediate adoption. I refer to the belief that, supposing such a work were erected, the lake would quickly be silted up by tailings derived from the diggings, so that the bottom being raised the weir would speedily become useless. Now the area proposed to be occupied by the reservoir is presently about as much exposed to those deposits as it would be then, and though some parts are so acted upon to a considerable extent, yet had the evil been of such proportions as to be practically felt, a necessity would have existed ere this for the immediate erection of a weir at the outlet, to counteract the shoaling process, and thus prevent a more rapid discharge of the water than would be consistent with its natural condition. An examination of the locality, however, would convince anyone that there is but little to fear from this evil assuming dangerous proportions; for, taking the Naseby diggings alone, it will be seen that even after about nine years of extensive sluicing operations, during which the heaviest flood on record has been experienced, the greatest distance to which even isolated shingle of large size, or small boulders, have been carried has been about three miles, and they are still about fifteen miles from the lake, and with less chance of making even the same progress again, for the creek becomes flatter as it approaches the low grounds. It will be observed, also, that until the whole lake, up to its lowest water-level, has been filled by solid material, its utility for storing flood waters is unimpaired. I am not in a position to state the contents to that level, but taking its depth at five feet— which I believe to be within the mark—it will be granted, I think, that even with much increased diggings it is safe for many years. If not, then the sooner the outlet is raised artificially the better.
But a real argument for a portion, at least, of this work is to be found in the neighbourhood—from the fact that there are now two outlets from the lake proper, while, before the 1868 floods, there was only one; and also the narrow gorge at the foot-bridge was widened by about an eighth part in the flood of 1870, and from the nature of the strata—being basalt, with very many joints, overlying clay—it is liable to greater extension, and, consequently, to allow the water to come more quickly towards the lower parts and facilitate floods. If it should so happen, in succeeding floods, that the same enlargement of these three outlets should continue, the utility of the lake as a regulating reservoir will be very much reduced; and the more rapid delivery of its waters may almost enable a flood equally as destructive as that of February, 1868, to result from less rain.