PROTECTION FROM FLOODS.

Otago Daily Times, Issue 21266, 21 February 1931, Page 8

PROTECTION FROM FLOODS.

REPORT ON LEITH AND LINDSAY’S CREEK. WORKS COSTING £617,538 RECOMMENDED. EXTENSIVE ALTERATIONS BELOW ST. DAVID STREET. WIDENING OF THE CHANNEL.

Works estimated to cost £617,538 are recommended by the Commission of Engineers, consisting of Mr F. W. Mac Lean (of Wellington), Mr J. M‘G. Wilkie (engineer to the Otago Harbour Board), and Mr J. G. Alexander (city engineer), which was set up nearly two years ago to report on floods in the Water of Leith and necessary protective works. After reviewing past floods the engineers express the opinion that the provision of works to prevent damage should be restricted as far as possible to residential areas where the effects are most serious. To provide works which would give immunity from damage in the case of such floods as that of 1929 in the upper reaches of the stream is in many places quite impossible, and where possible the cost would be out of proportion to the likely damage. The report, however, stresses the necessity of safeguarding the city water supply. A number of bridges are stated to have insufficient waterway, and the report indicates that substantial alterations will be necessary. Complete reconstruction is stated to be necessary at the King street bridge, and the engineers recommend a deviation of the stream from the sharp bend at Willowbank below George street to the bend below the King street bridge, and the construction of a new bridge practically at the entrance to the Botanic Gardens. Dealing with Lindsay’s Creek the report recommends the renewal of a number of bridges. At the point where the Main North road crosses the creek there is insufficient waterway under the bridge, and, owing to the nature of the creek, the construction of a new channel a short distance southward is considered necessary. Deviations and widening are also recommended. Returning to the Leith Stream the report states that a new bridge at Dundas street is necessary. From this point walls, widening, and the construction of weirs will be required. The Union street bridge has insufficient waterway, and the construction of a large culvert on the right bank is proposed. It is also considered necessary to widen the channel gradually between Union street and Leith street to 100 feet, and to continue this width to the outlet. A traffic bridge at Leith street and footbridges at Clyde street and Harbour terrace are proposed. An important recommendation involves the renewal of the Forth street bridge, the low level of which offers serious difficulties. To provide additional waterway at the Anzae avenue bridge concrete culverts on either side of the channel are proposed. The railway bridge will require to be made about twice its present length, with provision for subway vehicular traffic on either side. It is proposed that the bridge across the reclamation channel should be removed and a new bridge provided when necessary. Alterations in the bed of the stream from St. David street to the outlet are involved. The work, the report states, would, . if it were approved, be carried out only gradually. The total of estimated expenditure is composed of £480,135 with respect to the Water of Leith and £137,403 with respect to Lindsay’s Creek, nearly £200,000 being involved in the work below Union street. Boulder traps above George street require the outlay of considerable sums.

THE WATER OF LEITH. From its source at the saddle between Mount Cargill and Swampy Hill to its cutlet in Otago Harbour the Water of Leith has a course of slightly over seven miles, with a catchment area (excluding Lindsay’s Creek) of 7300 acres, or; 11.41 sauare miles, and, with its principal tributary, Lindsay’s Creek, a total area of 10,390 acres, or 16.24 square miles. The total fall in the stream is approximately 1200 feet, which may be roughly apportioned thus: To half-mile .. - - •• 300 feet Half-mile to two miles .. 200 feet Two miles to four miles .. 500 feet Four miles to seven miles 200 teet The lowest portion —the artificial channel across the Pelichet Bay reclamation, about 2200 feet-has a very flat gradient. It will be seen that for the greater part of its course the stream has an exceedingly rapid fall, and even in the lower portion, except in the artificial channel, the fall is very appreciable, tor a considerable distance the Leith flows in a narrow valley—practically a ravine—with steep slopes rising to the hills— Mount Cargill, 2232 feet, to the eastward; Flagstaff Hill, 2192 feet and Swampy Spur, 2185 feet, to the westward, ihe slopes are specially steep on the western side. There are numerous tributary streams the principal of which are West Branch, Morrison’s, Cedar, Nichol s, and Ross Creeks on the right bank; Cargill, Pine Hill, and Lindsay’s Creeks on the left bank. Some of these tributary streams rise at elevations of about 1800 feet above sea level. They have all rapid, in some cases precipitous, falls. In times of flood both the main stream and the feeder streams carry large quantities of trees, stone, and other detritus. Lindsay’s Creek, the largest tributary, which joins the Water of Leith in the lower reaches, will be referred to separately. The Water of Leith is of the nature ef a mountain torrent, and is subject, a,s such streams are, to sudden extreme floods. The compact form of its catchment area, no part of which is any great distance from the, critical flooding points, accentuates the rapidity with which the storm waters reach the main stream. In such catchments the occurrence of extreme floods depends rather on the intensity of the rainfall than on its tot*rl amount. LINDSAY’S CREEK. Lindsay’s Creek, the main tributary of the Water of Leith, rises on the upper slopes of Mount Cargill, at an elevation of approximately 1700 feet above the sea, and has a course of about four and a-quarter miles to its junction with the main stream at the Botanic Gardens, roughly, a mile and a-quarter from the outlet of the Water of Leith. The catchment area of Lindsay’s Creek is 3090 acres, or 4.83 square miles. In the first mile of this stream there is a fall of about 1200 feet, and in the following mile and a-half a further fall of 300 feet. Thereafter the stream has a moderate fall. It will thus be seen that for about two miles and a-half this stream also partakes in a marked degree of the character of a mountain torrent, carrying in times of flood considerable quantities of stone and detritus.

The slopes are generally much less steep than in the case of the Water of Leith; but, on the other hand, in the lower portion of the catchment area, the bottom and side slopes of the valley are occupied to a large and increasing degree by streets and buildings, intensifying in this way the rapidity with which storm waters reach the main stream. As in the case of the Water of Leith, the catchment area is of a compact character. As already indicated, the Water of Leith and its tributaries are very liable, from the character of their catchment areas, to sudden extreme floods. Before the areas in the lower and flatter portions of their valleys were built over, the streams were free in flood time to overflow on to the adjoining lands, and to change the course of their channels. With the occupation of the land by streets and buildings, and the reclamation at the mouth of the Leith, this freedom has been restricted. Storms causing floods usually occur during southeasterly winds, generally accompanied by iow barometer readings and very high tides in the harbour. It is to be noted that some of the land built on near the harbour is very little above the level of these high tides. All these above-mentioned conditions naturally affect the difficulty and expense of dealing with flood water. FORMER FLOODS. That floods in the Water of the Leith and Lindsay’s Creek have been of fairly frequent occurrence, and that some of these floods were of considerable magnitude, is apparent from the reports in the Otago Daily Times, but the actual amount of flood discharge can only be a matter of conjecture. It is important to note, however, that in most cases the floods were “sudden,” and were therefore evidently due to rainstorms of high intensity rather than to long continued rainfall. The worst of these earlier floods of which we have reports seem to have been those in February. 1868, February, 1877, and November, 188.3. On these \Ve offer a few commentsFlood of February, 1868. —Considerable damage occurred in the lower reaches of the Leith, and some bridges were destroyed. A large portion—“ about an acre ” —of the Botanic Gardena was swept away. The rainfall recorded for the month of January was 8.08 inches and for February 5.07 inchesFlood of February, 1877. —This appears to have been the heaviest of those re* ported. The flood brought down very large trees. Dunedin was “practically shut off from the surrounding country.” The lower part of Frederick street was “ flooded to an extent never known before.” The Botanic Gardens again suffered severely, and a protective breastwork constructed after previous floods was carried away. A small house, near St. David street bridge, was swept away Low lying sections in North-East Valley were covered with water, and numerous residences suffered considerably. The water works overflowed considerably, and “a body of water about four feet deep was sweeping over the banks of the reservoir.” If anything like this last occurred, the flood must have approached in magnitude, if it did not exceed, even

that of March, 1929. The rainfall recorded for the month was in January, 3.94 inches, and in February 6.24 inches. Flood of November, ‘ 1883. —Bridges in. the neighbourhood of Woodhaugh were severely damaged and rendered unsafe. The banks of the stream were undermined. About a mile above the Paper Mills, a length of the roadway was swept away bodily, and the stream made a new course for nearly a quarter of a mile. Great damage was done in the North-East Valley. “ The inundation occurred within little more than a quarter of an hour.” The rainfall recorded for, the month of October was 2.05 inches, and for November. 5.01 inches. RECENT FLOODS. ■ The floods which have occurred in recent years have been more or less carefully investigated, and a study of these is therefore an aid in forming conclusions as to what is required in provision for and prevention of damage from future floods. Flood of December 24, 1911: This flood appears to have been due principally to heavy rainfall over a portion of the catchment area. Most of the rain gauges showed a small rainfall, while that at Opoho (Mr Paulin, observer) registered about 12 inches in one hour. There was a heavy flood in Pine Hill Creek. The flood discharge was estimated to’ be about 3000 cubic feet per second in the Water of Leith at George street bridge, with “ probably 1000 cusecs in Lindsay’s Creek,” making 4000 cusecs below that point. It was considered that there was a general lowering of the bed of the Leith through the town by as much as 18 inches. Considerable damage was done by overflows in the lower reaches of the stream.

Flood of August 13, 1913: This was a minor flood, an overflow taking place at Harbour terrace, due to some extent to the work then being carried out by the Harbour Board in connection with the new reclamation channel. The rainfall during 24 hours, ended 9 a.m. on August 14, at Ross Creek reservoir, was 2.88 inches, and the flood discharge was estimated at about. 800 cusecs.

Flood of June 20, 1926: The rainfall recorded for 24 hours was 2.45 inches at Opoho (Mr Paulin), 2.04 inches at Kew (Captain Easther). and 2.31 inches at the Musselburgh Pumping Station. Following on the flood which resulted from the rainfall, there was a difference of opinion as to the works which had been carried out, notably the provision of the series of weirs in the bed of the channel below Union street, and as to the capacity of the reclamation channel to carry flood waters. The matter was referred to the Public Works Department, and a report obtained on the flooding of the Water of Leith and on the works carried out. Certain remedial measures were recommended (and carried out) to prevent damage by similar floods ranging up to 5000 cusecs, but it was pointed out that no margin of safety would be left if a greater flood were experienced. Flood of April 22 and 23, 1923: This flood was of a_ very extensive character, and caused serious damage. The rainfall recorded at various stations was as follows: Pumping Station. Mussel- Botanic Whnro Sullivan burgh. Gardens. Flat. Dam. 2i hours: Inches. Inches. Indies. Inches. April 22 .7110 1.210 I.SSO 1.125 ■ April 23 9.020 G.SIO 7.700 9.375 These records were obtained from ordinary rain gauges, automatic rain gauges not having been then installed, but the rainfall at Sullivan dam was measured at intervals, thus:— 8.0 a.m. to 12 noon .. .. 1.500 inches 12.0 noon to 6.0 p.m 2.125 inches 6.0 p.m. to 6.0 a.m. .. ~ 5.750 inches Total .. 9.375 inches The average intensity from 6 p.m. on April 22 to 6 a.m. on April 23 was thus nearly half an inch per hour, but it is probable that the 'maximum intensity must have exceeded this greatly. During the evening of April 22 the velocity of the current was measured (by the harbour engineer) below Harbour terrace, and found to be nearly 18 feet per second (say, miles an hour), representing a discharge of over 3000 cusecs, but the flood had not then reached its highest point. The flood discharge, very carefully investigated, was estimated at not less than 7000 cusecs. A large proportion of this overflowed, flooding the low-lying areas", and causing a great deal of damage. Flood of May 13, 1923. —Following on the flood of the previous month, this was of considerably less intensity. The rainfall on this occasion was as follows: t , Pumping Station, Mussel- Botanic IVhare Sullivan burgh. Gardens. Flat. Dam. 21 hours: Inches. Inches. Inches. Inches. May IS .. .70 1.05 1.95 t 1.25 May 11 .. 1.32 1.20 2.90 3.19 The rain fell on May 13 for five and a-half hours, and on May 14 for eight hours. The average intensity was thus under half, an inch per hour. The flood discharge was estimated at about 4000 cusecs. The April flood brought down a large quantity of boulders and shingle which lodged in the channel in the neighbourhood of Harbour terrace, and had been only partially removed before the May flood occurred, thus restritcing the waterway and causing the stream to overflow and do damage. m . Flood of March 19 and 20, 1929—This flood is believed to be the highest experienced in the Water of Leith, and will therefore be commented on at greater length in its various aspects. THE RAINSTORM. The storm which caused the floods was a cyclone which crossed the southern portion of New Zealand. There was general rain in connection with this storm, but the heavy falls were in the eastern district from about Balcmtha to Christchurch, while the greatest intensity was confined to the district from a little south of Dunedin to just north of Oamaiu. Although cyclones are usually or rare occurrence in New Zealand, they are fairly frequent in some years. This had been the case for 1928-1929. The_ general movement of the disturbance is nearly always from the westward, the heaviest rainfalls usually occurring with southeasterly winds. This (south-easterly) was the direction of the wind on March 19, 1929. THE RAINFALL. Rain gauges are established at_ a considerable number of points in the neighbourhood of Dunedin, and the records from these have been of great value in the investigation of the Hoocl. This is specially the case in the records of the automatic rain gauges established since the flood oi 1923. Comparatively few records, however, are available within the catchment area of the Water of Leith (four), and of Lindsay’s Creek (one), and that of these only one (Maori Hill reservoir, at the extreme southern boundary of the .Leith area)is from an automatic ram gauge. A very careful examination ot the Silverstream and its tributaries, and of Burns Creek and Williams Creek (on the Waitati side of Swampy Spur), after the flood, convinced us that the intensity ox the rainfall, judging by the effects of the storm, had been very similar in the case of Ross, NiehoTs, Cedar and Morrisons Creeks, and in west branch, to that in the Silverstream and Burns Creek areas. It is interesting to note that although there was no marked difference in the rainfalls taken altogether during the 24 hours in 1923 and 1929, the consequent flood in 1929 was tar in excess of that in 1923. The cause of this is clear from the records of intensity of the rainfall shown by the automatic rain gauges (established since the 1923 flood). Figures show very emphatically that in a small catchment area of the character of that of the Water of Leith, the in tensity of the rainfall is the essential factor, and also show the value of automatic recording rain gauges. The ratio of the rainfall in one hour to the average per hour over the whole day (the latter being usually the only record available) varies in this instance from nearly four times as great in the case of Burns Creek to a little over tuice as much at Maori Hill Reservoir. Taking the stations at which the maximum intensity could be judged, the average ratio was over three times and a-quarter If records of intensity bad been available for the Ross Creek station, it is probable that this ratio would have been exceeded. It has to be recognised further that in such a case as that of the Water of Leith catchment area, where storm waters reach the lower portions of the stream quickly, an increase in the intensity of the rainfall, even for a limited time, following on or during continuous rain of varying but great intensity, lias a very marked effect on the magnitude of a flood and on its consequences. The records show that after continued rainfall during the early part of March 19, the storm broke with great intensity during the afternoon, lasting till a late hour, and. in the case of the Burns Creek area, till the early

hours of March 20, the greatest effect being experienced on the FlngstaffSwampy Hill ridge, the eastern side of which discharges into the Leith. FLOOD DISCHARGE. In flood investigations, it is essential that the flood discharge, or maximum quantity of water flowing at any time during the flood, should be ascertained with as great a degree of accuracy as possible. It has been pointed out in dealing with former floods that no estimate appears to have been made of the flood discharge in these. In more recent floods the investigations made indicated a flood discharge ranging from 4000 cusecs in the 1911 flood to not less than 7000 cusecs in the flood of April, 1923. Our preliminary investigations into the flood of March, 1929, indicated that the amount of the flood discharge was far in excess of anything of which a definite record could be obtained. To place this all-important point beyond question as far as possible, it was considered necessary to haye a detailed investigation made into this portion of the problem by a qualified man who could devote the necessary time to the work before the traces of the flood were oblitr erated. For this purpose, Mr R. a. Cree Brown, 8.E., A.M.Inst.C.E. (possessing special qualifications for the work) was engaged to make a thorough investigation of the flood discharge in the main stream and in the principal tributary streams . The. recognised methods of ascertaining flood discharge, in order of reliability, are:—

1. By cross-sections of the stream channel and measured velocities of the current. , 2. By cross-sections of the stream channel with calculations based on factors ascertained from exhaustive experiments and on experience. 3. By the amount of rainfall as shown by rain gauges in the catchment area, and assumed percentage of “ run-off i.e., the proportion of the rainfall reaching the streams. This assumption has to be based on experience in similar conditions. 4. By inference from ascertained or estimated flood discharges in similar localities. With reference to these methods, we

may remark: — 1. This method could not be adopted as no measurements of the velocity could be made. The highest point of the flood occurred during darkness. 2. This is the method adopted tor tne purpose of this report. . 3. The difficulty in applying the third method to this case is that dependence has to be placed on records from rain gauges, some of which are outside the actual catchment area, although in fairly close proximity to it, and that, of those within the catchment only one is of the automatic type. As previously stated, however, a careful inspection of the Silverstream and Waitati catchments showed that the conditions, so far as could be judged by the effects of the storm, were very similar to those in the Leith catchment. In the Silverstream and Waitati catchments, there are two; automatic recording gauges and one ordinary gauge which was read at intervals during the storm. The third method has, therefore, been used with some degree of confidence as a check on the calculations made by the second method, and with a fairly close agreement in the results obtained. 4 Further reference will be made to the fourth method when dealing with the problem of future floods. It will be seen from Mr Cree Brown s report that he estimates the flood discharge in the Water of Leith and Lindsay’s Creek, respectively, as:— ‘Water of Leith above Lindsay’s Creek, 8900 cusecs.

Lindsay’s Creek, 2440 cusecs, Water of Leith below Lindsays Creek, 11,340 cusecs. On the basis of the area of the catchments, this works out at: Cusecs. Cusecs. Per acre Per sq mile Water of Leith, above Lindsay’s corner .. 1.22 or 780 Lindsay’s Creek 79 or 605 Water of Leith below Lindsay’s Creek .. 1.09 or 700

The figures for discharge have been checked by method 3 —i.e., by comparison with the run-off deduced from the isohyeThese charts were prepared carefully from the rain gauge records, and we believe they represent with reasonable accu-racy-the distribution and rates of rainfall in the catchment during the storm. We have also assumed that the whole of thesfe hourly rainfalls (i.e., 100 per cent, of “ run-off), during the greatest intensity of the storm, reached the streams. This assumption is based on the facts that previous rain had saturated the ground (which is not generally of a pervious character), and that the intense rainfall lasted for a considerable time — longer than was necessary for the storm waters to reach the streams. On this basis, the total flood discharge was found to be practically the same as arrived at by the previous method. The quantities stated in cubic feet per second may be more readily appreciated if we give the discharge in well-known rivers:—

Waitaki River.—The normal summer flow of this river at Kurow is from 12,000 to 14,000 cusecs. The winter flow is from 4000 to 5000 cusecs, and has been as low as (1930) 3000 cusecs.

Taieri Eiver,—The average flow at Outram is from 2000 to 3000 cusecs. lu dry seasons the flow there has been as low as a few hundred cusecs, while an “ ordinary ’’ flood is about 20,000 cusecs. The maximum flood investigated was computed at 50,000 cusecs. Clutha Eiver.—The ordinary flow at Balclutha is about 20,000 cusecs. The great flood of 1878 was estimated at 180.000 cusecs, and that of 1919 at 120,000 cusecs. Kawarau Falls (Lake Wakatipu).—The flow varies from 3000 cusecs in winter to 5000 cusecs in summer. The average flow throughout the year is approximately 4000 cusecs. Waikato River. —At the outlet from Lake Taupo the average flow in the Waikato River is 6000 cusecs. A high flood measures 10,000 cusecs. At Arapuni the flow varies from a minimum of 5000 cusecs to a maximum of 32.000 cusecs. The average flow is about 7000 cusecs. EFFECTS OF THE FLOOD. Water of Leith (upper reaches). —The intercepting channel around the northern and western sides of Sullivan dam overflowed (as in the 1923 flood), the excess passing into the dam. At an early stage in the storm the supply from Burns, and Williams Creeks (in the Waitati catchment) was stopped by the carrying away of the headwords and pipes there. The total quantity of water drawn from these creeks is less than four cusecs —an insignificant quantity in relation to that derived from the rainfall during the storm. Some scouring out of the ground near the overflow channel —some distance below the dam embankment—took place. There was also a slip in the gully below the dam which blocked a drainage culvert. The dam itself suffered no damage.-. There seems to have been an impression that the “sluice gates” at the dam were opened to allow the water in the dam ' to escape in order to safeguard the works, and that the flooding lower down wae due partly to this. It is difficult to understand how such an impression could originate, as there arc no sluice crates. The discharge from the dam to the water supply pipes is by syphon, which could not exceed its normal capacity. The only other outlet is by the overflow channel already referred to, which was designed to carry off surplus water, and functioned accordingly, there was a considerable overflow of the stream at M'Kcnna’s bridge, damaging the roadway owing to the insufficient waterway at the bridge. Below the junction of the west branch there was an extensive deposit of boulders, etc., brought down from the west branch, and the water supply main was undermined. At the scenic reserve ” considerable erosion took place. The concrete bridge carrying the road near Mr Booth’s farm was undermined and cracked. At Flett’s bridge an overflow damaged the roadway. At Cedar Creek the Leith Valley road was washed away A portion of the schoolhouse ground was flooded, and considerable erosion took place at the bank of the stream adjoining the playground. In addition to the above, a great deal of damage was done by the undermining of walls and other protective works, roadway washed away entirely in places, and numerous slips. The 12-inch water mam was also injured at various points. Middle Reaches: Harbour’s Bridge to Lindsay’s Creek. —“ Harbour’s ” bridge proved quite inadequate to carry the flood waters, and very considerable damage was done in the undermining and destruction of protective walls, erosion of banks, etc., particularly below the bridge. A very extensive deposit of boulders and detritus in the bed of the stream from the sawmill to near the “ Rockside ” bridge forced the

stream into new channels. Below the paper mills, there was another very large deposit of boulders and detritus. The bridge leading to the Woodhaugh quarry was carried away. At and near Malvern street bridge the stream overflowed. Between Malvern street bridge and George street bridge the Public Gardens were flooded to some extent, but no great damage, was done. The waterway provided by the George street bridge was totally inadequate for the flood waters. The water was banked up by the bridge, and was thus acting under a 'considerable “ head." The water which could not pass through the bridge escaped along the streets, rejoining the stream lower down. Some idea of the strength of the current at the place may be judged by the fact of a wave overtopping the iron hand rail of the bridge. This was vouched for by an experienced engineer observer (Mr T. G. G. Beck),

Owing to the obstruction to the passage of the flood waters, partly due to the sharp bends in the stream channel, partly by the very inadequate waterway in the King street bridge, a very large body of water overflowed below the George street bridge towards the Botanic Gardens and Opoho road, along the North road, joined the Lindsay’s Creek flood waters, and passed by way of Lindsay’s Creek back to the main stream of the Leith. Tributary Streams. —At West Branch the waterworks intake weir was undermined, and the branch water main was injured. The flood in this stream carried with it large quantities of boulders and detritus, which lodged in the Leith Valley. Morrison’s Creek and Cedar- Creek showed iii a marked degree the effects of the storm, boulders, trees, and detritus being carried by the flood waters in large quantities. The breaking away of water, temporarily impounded by numerous slips, aggravated the effects of the flood in the main stream.

The branch water supply from ■ Morrison’s Creek suffered very severely, and a length of about 500 fet of the pipe line was swept away, the greater part of the remaining pipe line being damaged or endangered, and the total length of 2343 feet had to be relaid on a safer alignment. This, with the cutting off of the supply from Burns Creek and Williams Creek, and from the West Branch, together with the injury to the water main from Sullivan dam, seriously affected the city water supply. Although not directly connected with the present investigation, but at the same • time affecting consideration of the subject, it is well to note that the damage to the Silverstream water race, caused simultaneously by_ the same storm, curtailed further the city water supply, which was thus completely disorganised. The flood in the Silverstream was very great, higher *at “ Hazlett’s ” bridge in the Taieri Plain than the 1923 flood by approximately five feet,

At Ross Creek the by-pass, or storm water,. channel at the reservoir proved quite inadequate, and the flood waters overflowed into the reservoir. The upper or settling basin became overcharged, and the water overtopped the dividing dam by a few. inches. The overflow weir in the main dam was fully taxed, but was sufficient to protect the dam. The fine silt brought down by the flood waters caused considerable discoloration of the water in the dam, thus further affecting the city water supply. A good deal of damage was done by the flood above the reservoir. In the Lindsay’s Creek area the average intensity of the rainfall was evidently less than in the Water of Leith area. The movement of boulders, which was such an important feature of the flood in the latter, was confined principally to the very steep upper reaches, few being carried further than the upper end of the open valley, but the flood fvaters came down in great volume. The bed of the stream being generally, as far as the crossing of the Main North road, at a comparatively small depth below the “floor" of the Valley, there was a general flooding of the latter, A considerable body of water flowed along the Main road and other streets from the quarry downwards. Owing to the insufficient waterway at the bridge carrying the Main North road over the stream, and the constricted character of the channel in the neighbourhood, there was a very considerable overflow at this point, a large volume of water flowing into the Main road, greatly augmented by the overflow coming from the Water of Leith below the George street bridge. As already stated, there was general flooding of the lower levels of the Valley, and the Botanic Gardens were flooded over a fairly large area. Lower Heaches Lindsay’s Creek Junetionto Outlet.-—This portion of the Water of Leith naturally suffered most from the flood, being subjected to the combined volume of the storm waters in the Water of Leith and Lindsay’s Creek, and much of the adjoining land being at a low level. The bridges at Leith street, Clyde street and Harbour terrace were swept away by the flood, the bridges at Castle street and Dundas street were damaged, waffs were undermined, and other protective works were seriously damaged. About 500 houses in the north end of the city were flooded to depths ranging from a few inches to four feet. A great volume of water overflowed—practically a flood in itself—and swept along the streets on the lower levels ns far as, and beyond. Lower Rattray . street. This overflow took place principally in the part of the channel between Harbour terrace and the bend below the Union street bridge. The conditions were greatly aggravated by the very large quantity of boulders, tree trunks and detritus brought down by the flood, blocking the waterway in the channel. , , , , Some 7000 cubic yards of silt and detritus brought down by the flood waters were removed from the city streets, in addition to about 1200 cubic yards removed from private properties. ihe quantity of boulders, shingle and silt afterwards removed from the channel above and below the paper mills and throughout the lower portions of the stream amounted to 33,000 cubic yards. As an indication of the velocity of the current, boulders up to about Two feet in diameter were brought down ,by the flood waters to Harbour terrace. Altogether, the effects of the flood were disastrous, causing a great deal ot damage to property, but, fortunately, without loss of life. Up to this point in this report reference has been made to past floods in the Water of Leith in an -endeavour to ascertain the extent to which the locality has been subject to flooding, as a guide to what may be anticipated in the future. FUTURE FLOODS. The question of the amount of flood discharge which should be provided for in future has given us much thought. The greatest flood which has taken place in the Water of Leith is apparently that of March, 1929, although other floods—e.g., that of 1877—may have been of similar magnitude. Taking the whole evidence into consideration, we are of opinion that the 'maximum discharge was approximately 11,500 cusecs. „ -n -tr u The rainfall in the North-East Valley (Lindsay’s Creek area) was less intense than in the Water of Leith area. A comparatively small variation in the course of the cyclone might have resulted in a considerably heavier rainfall in the NorthEast Valley. It has been pointed out that a large portion of that area is occupied by streets and buildings. On this account rainfalls reach the streams quickly. This occupation of the area is likely to increase largely. For these reasons we have come to the conclusion that not less than 14,000 cubic feet per second should be provided for, oil the basis of past experience. It is necessary, however, to consider not only past floods, and whether there appears to bo any likelihood of a recurrence of equal floods, but also those which have occurred in other localities where the conditions may be considered more or less similar to those in the Water of Leith catchment area. This is the fourth method of estimating flood discharge referred to previously in this report. - Until quite recently there has been no systematic investigation into rainfalls throughout New Zealand. The recent investigations have been of a general character, and, though exceedingly valuable, do not give all the information required in such an Investigation as the present. We have made repeated reference to the intensity of the rainfall as the essential factor in dealing with floods in the Water of Leith. The available records throughout New Zealand of measured intensities from automatic rain gauges are very limited. Where daily records of rainfall only are available, the intensities must necessarily be estimated.

Startling as the amount of the Hood discharge in the Water of Leith in the flood of March. 1929, may appear to be, it can bo seen from examples of rainfall and flood discharge in other parts of New Zealand that there is a possibility of even heavier rainfalls and larger floods occurring.

After the most careful consideration of I the whole of the surrounding circumstances in such cases as we have studied, i with due regard to the probably infrequent occurrence of such catastrophic floods, and the huge expenditure that would be involved in preparing for them, I we are of opinion that provision for pro- ; tection against damage from floods in the j Water of Leith greater than 14,000 cusecs. i is not warranted.

REMEDIAL MEASURES. In considering this part of the problem it has appeared to us that the provision of works to prevent damage from floods should be restricted, as far as possible, to the residential areas where the effects are most serious.

To provide works which would give immunity from damage from such a flood as that of March, 1929, in the upper reaches of these streams is in some places quite impracticable, and, where possible, the cost would be out of proportion to the damage likely to result. In the case of the Water of Leith, it is certainly essential to safeguard in every possible way the city water supply. We have assumed that a flood discharge of 14,000 cuoic feet per second in the lower part of the Water of Leith (below the junction of Lindsays Creek), with proportional discharges in the higher portions, has to be provided for. In the proposed works we nave endeavoured to limit the velocity of flood waters to 16 feet per second, especially in the lower reaches, though this has not always been possible. The first_ essential is the prevention, ns far as possible,_ of the passage of boulders and other detritus likely to obstruct sen ously the flow of storm waters. It is obviously of little avail to provide a channel of dimensions necessary to carry the flood discharge if that channel is likely to be constricted by the deposit of large quantities of boulders, and to provide a channel large enough to be unaffected in this way is impracticable. In Lindsay’s Creek this question is not serious. The quantity of boulders carried down from the very steep upper reaches is_ comparatively limited, and can be dealt with without special works. In the Water of Leith, large quantities of boulders and detritus are brought down by floods, and,'unfortunately, the general gradient of the stream, without any fairly level reaches, precludes any thoroughly efficient settling ground. At various points in the course of the stream the conditions favour the deposit of stone to a limited extent, and it is desirable to encourage this where the stone can be utilised and the deposit will not cause damage. It was at one time proposed to form a “ boulder trap ” in the lower portion of the Leith—near the Leith street bridge—but it is highly desirable that boulders should be intercepted before they reach this point. The only positions available for the reception of large quantities of boulders and detritus are between Harbour’s and Rockside road bridges; and bejow the junction of Ross Creek. At the latter place land was previously purchased for the _ purpose, and has the advantage of proximity to the stone crushing plant. Great quantities of boulders and detritus were deposited at these two places during the flood of March, 1929. There is an objection to the upper of these proposed boulder traps in that it is a considerable distance from the existing stone crushing plant, but this might be overcome to some extent by special methods of transport, or by the installation of a separate crushing plant. The stone brought down is of excellent quality, but the handling will add to the cost of the crushed material, and considerable quantities of detritus which cannot be utilised profitably will have to be removed. For these boulder traps to be effective, the deposits will have to be removed regularly. With these two boulder traps and the regulation of the current velocities, it is reasonable to anticipate that comparatively few boulders_ and other large detritus would be carried down to the lower reaches. . THE PROPOSED WORKS. WATER OF LEITH. SOURCE TO HARBOUR’S BRIDGE LOCALITY. In this portion of the stream it is proposed to restrict the remedial measures to the safeguarding of the water supply works, to the provision of sufficient waterway in the bridges for the passage of the flood discharge adopted, to the repair and strengthening of protective works already provided, and to clearing the waterway generally. Water Supply.—Where the water main has been endangered (e.g., below the junction of the West Branch), or likely to suffer damage, it is proposed to remove the pipes to a safer position, where practicable, and where this cannot be done to provide protection. At Morrison’s Creek a large amount of work has already been carried out,'and the supplementary water supply there has been rendered as safe from injury as possible. At Ross Creek reservoir, where the bypass channel proved altogether inadequate, a new channel on the southern side of the reservoir is being constructed. The damage to the works above ' the reservoir has been _ repaired. The chief damage likely to arise in the upper reaches of the Leith from very high floods is in the blocking of the road by washouts and slips. Bridges.—ln most of these there is insufficient waterway. A new bridge is required below M'Kenna’s. It is proposed to provide a new channel to cut off the abrupt bend in the existing channel, and construct the new bridge in the line of the new channel. The bridge near the road to Booth’s farm was damaged by the flood of 1929, and a temporary bridge was constructed. A new bridge, immediately above this temporary bridge, with single waterway and with improved approach, is proposed. The waterway in Stonyford bridge is proposed to be increased by deepening the invert. This will involve the altering of the stream channel for some distance above and below the bridge. It would be advisable to increase the waterway at Flett’s bridge by deepening the bottom at the upstream side; a bank is required on the road side above the bridge, and the wing wall at the southwest corner should be raised to prevent overflow- A considerable amount of work will have to be carried out in the repair, strengthening, and renewal of protective works (a large amount has already been done) in preventing erosion and m clearing away obstructions. At the school, protective works are required to prevent further erosion. HARBOUR’S BRIDGE TO LINDSAY’S . CREEK JUNCTION. The channel of the stream, for some distance above Harbour’s bridge will require to be altered to give a better course and direction for flood waters. The banks also require protection. . Harbour’s bridge has very inadequate waterway, and, owing to the “ set of the current, part of that waterway was blocked by the deposit of boulders and gravel. The bridge requires rebuilding and enlargement. ’ ~ From the bridge downstream to Davidson’s mill the retaining walls on the left bank require complete renewal and altered alignment. The right bank requires protective pitched slopes, and a portion of the bluff opposite the mill should be removed. . , , , Below the null it is proposed to form the upper of the “ boulder traps previously referred to. This involves the removal of the existing large deposit, protective walling along the left bank of Bine Hill Creek, protection of the right bank, and the construction of a low weir across the bed of the stream. These protective works require to be extended as far as the Rockside road bridge, and the channel of the stream reformed. The Rockside road bridge will require to be reconstructed with increased waterway, and from this point to the pipe bridge it will be necessary to reconstruct the channel and to protect the bottom and sides, making due provision for the supply of water to the paper mills. Below Ross Creek junction the right bank has already been protected by a concrete wall. Walls are required along the left bank for a similar distance, and the bed of the stream requires regrading and protection from scouring. Immediately below this a new bridge (to replace that carried away by the 1929 flood) is required to give access to the quarry and stone-crushing plant. From this point to the eastern end of the Woodhaugh street sections, it is proposed to establish the lower of the two “ boulder traps ” previously referred to, by the construction of a moderate height of weir at the lower end, the removal of the existing deposit of boulders and detritus (except what is required to form a protective bank along the back of the Woodhaugh street sections), and the excavation of as wide an area as possible.

To provide the increased waterway required at Malvern street bridge, ij; is proposed to construct a new bridge of

single span, and deepen the channel. The latter involves the re-alignment of the sewer which crosses the stream at this point. Above the bridge, new walls will be and suitable weirs to protect the bed of the stream. From Malvern street to George street it will be necessary to widen the channel and protect the right bank in places. To regulate the velocity of the current, a series of weirs will be required. At George street bridge it is proposed to provide the greatly increased waterway required by deepening the channel, underpinning the foundations of the existing structure, and providing a substantial concrete invert. Below George street the sharp bends in the channel and very inadequate waterway in King street bridge cause serious obstruction to the passage of flood waters. Without complete reconstruction, the necessary waterway cannot be provided at King street bridge. We are of opinion that the most satisfactory solution is to make a deviation of the stream from the sharp bend at Willowbank below George street to the bend below King street bridge, and to construct a new bridge practically at the entrance to the Botanic Gardens. Only a small portion of the ' gardens would be affected, involving: a rearrangement of the entrance, paths, and flower beds. A much more extensive deviation from the bend at Willowbank to the junction of Lindsay’s Creek was considered, but was dismissed as involving very large expense and interference with the Gardens, without commensurate advantages. From the end of the proposed deviation to Lindsay’s Creek junction the channel will require widening and deepening, the strengthening of the existing walls, and the construction of new walling on both banks. A new footbridge at Castle street will be required. LINDSAY’S CREEK. It is not proposed at present that any works except of a minor character should be carried out northwards of Beechworth street, but, eventually, under circumstances arising in the future, such as closer settlement, this aspect of the matter will have to receive further consideration, together with the interception of boulders from the upper and steeper portions of the stream. From Beechworth street downstream for a considerable distance the bed of the stream is so little below the general level of the_ Valley and of the streets, that deepening of the bed and widening where necessary, together with walling, are required to provide adequate waterway. The renewal of a number of bridges will also be necessary. At the point where the Main North road crosses the stream (at the junction of Craighleith street), not only is there insufficient waterway in the bridge, but the channel of the stream for some distance below the bridge is so tortuous nni constricted that a considerable alteration is necessary for the improvement of the dischafge capacity of the stream channel. To deal with the difficulty, it will be necessary to construct a new channel and bridge a short distance southward. Considerable alteration to the sewers is involved. Deepening the channel will be necessary from this point throughout tho course of the stream to the junction with the Water of Leith,' with extensive walling and other protective work; alterations to sewers and water pipes will also be required. It is proposed to make a deviation of the stream where it passes through the property of the trustees of the late Mr Robert Glendining and adjoining owners. At Chambers street, a new bridge will be required, and at Opoho road, a large culvert to provide the necessary additional waterway will have to be constructed. Where it passes through the Botanic Gardens the stream will require considerable widening and deepening with protection of the sides and bed from erosion. LINDSAY’S CRREK JUNCTION TO OUTLET OF LEITH. Between Lindsay’s Creek and Dundas street, a considerable widening of the channel will be required. This will; have to be effected on the left bank to avoid interference with the properties on the right bank. The widening will involve some encroachment on the Botanical Gardens reserve and a slight reduction in the width of the short street. Gore place. Walling on the right bank and protective pitching of the slope on the left bank and weirs to regulate the current velocity will be required. We have considered the question of providing for a considerable portion of the flood discharge by an overflow tunnel from a little above Dundas street towards Harbour terrace, and thence by a channel to join the main stream near Anzac avenue. If this were feasible, the difficult problem of providing for the flood discharge between Dundas street and Anzac avenue would be greatly simplified. The cost of such a scheme, however, not only in the construction of the tunnel and channel, but in the purchase of property, renders the proposal impracticable. The position at Dundas street is affected by the Leith street sewer, which crosses the Water of Leith a short distance below the bridge. This bridge was seriously damaged by the 1929 flood, A new bridge with greatly increased waterway as required. From Dundas street to St. David street the width of the channel is restricted by Montgomery avenue on the right bank and the properties on the left bank. It is proposed to construct walls on each side for the whole distance, and form a fairly high weir at the sewer crossing to deepen the channel sufficiently to provide the necessary waterway. Weirs will be required to regulate the current velocity between Dundas street and St. David street. A new footbridge will be required at St. David street. Between St. David street and Union street, in front of the University buildings, the channel will be wider with vertical walls on the right bank, and grassed slope—pitched at bottom —on the left bank. Below St. David street the channel will be deepened. Union street bridge has, like others, insufficient waterway. It is proposed to provide the additional waterway required by the construction of a large culvert on the right bank. This will involve the widening of the channel on the right bank, near the old Dental School, and also on the south side of the bridge. Considerable alterations to sewers are also involved. At present those serving properties in Castle street connect with the Leith street sewer, by a sewer in Union street. This will have to be closed and a new sewer provided via Castle street and Albany street. Between Union street and Leith street the proposals include widening the channel gradually to 100 feet, constructing walls on each side, and providing for protection of the bed of the stream by means of weirs. This width (100 feet) of channel is to be continued to the outlet. In the 1929 flood the bridges at Leith street, Clyde street, and Harbour terrace were destroyed. It is proposed to provide a new traffic bridge at Leith street and footbridges at Clyde street ana Harbour terrace. , .... At Forth street a new bridge with increased waterway will be required. The low level of this bridge offers very serious difficulties in the way of a satisfactory scheme for the discharge of flood waters. It is, therefore,’ necessary to raise the level of the roadway at this bridge and its approaches between Albany strect-'and Union street, but no material disadvantage to traffic will be caused thereby. At Anzac avenue bridge it is proposed to provide for the necessary additional width of waterway by the construction of concrete culverts on each side of the channel. * ... The bridge across the reclamation channel, between Anzac avenue and the railway bridge, will be removed, and a now bridge provided when necessary. The railway bridge will require to be made about twice its present length, with provision for subway vehicular traffic on Alteration in the levels of the bed of the stream from St. David street to the outlet into the harbour is involved in these proposals. Deepening and re-grad-ing will be necessary, but it appears possible to avoid interference with the Leith street and Harbour terrace sewers which cross the channel. From Clyde street to the outlet, the invert of the channel will be concrete. Elsewhere, weirs will be constructed where necessary to reduce velocity and to protect the bed of the channel from erosion. A COMPLICATED PROBLEM. The problem of how best and most economically to afford reasonably effective protection against floods in the Water of Leith is a most complicated one. Although greater floods than those provided against by our proposals are not beyond the bounds of possibility, we consider that the likelihood of their occurrence is not such as to justify the greatly-increased expenditure in protective works which would he necessary to deal with such floods. We are convinced, however, from the records of past experience ia the Leith

and Lindsay’s Creek watersheds, and; from what has been experienced in otherlocalities, that it would be unwise to provide for smaller flood discharges than those assumed by us. We recognise that the cost of carrying out our proposals will be high. The work, however, if approved, would only be carried out gradually, and each part of it can be executed in accordance with, and as part of, the complete scheme.; We desire to express our thanks for, the valuable information given to us freely by Dr Kidson, Director of Meteorology, the officers of the Public WorksDepartment, the observers at the rain gauge stations with whom we have cor-, responded, and others. ESTIMATES OF COST. WATER OF LEITH. Harbour to upper side Anzac avenue £63,040 Upper side Anzac avenue to lower side Union street 124,035 Lower side Union street to upper side Dundas street .. .. 63,831; Upper side Dundas street to Lindsay’s Creek 23,258. Lindsay’s Creek to diversion below King street /. .. 17,104 Diversion channel at King street 26,603; Diversion channel to upper side, George street 11,995 Upper side George street to lower boulder trap below quarry .. 28,219 Boulder trap at quarry and up to Ross Creek 22,372 From Ross Creek to and including upper boulder trap •. .... 42,965 From upper boulder trap to end of special work above Harbour’s bridge 48,673 Remainder of channel 8,040 £480,135 LINDSAY’S CREEK. From Leith to upper end of diveraion above North road bridge £67,993’ From termination of above to Beechworth street 68,410 Minor works above Beechworth street 1,000 £137,403