The hydrology of the Lower Mekong provides a focal point for Peter Adamson as he reports on the potential impacts of developing hydro power in Yunnan, China

Regulation of the downstream flow regime, defined as some degree of de-naturalisation of the seasonal flow pattern by reallocating water from the wet to the dry season, is often perceived to be one of the adverse impacts of reservoir development. If the resource development does not involve the consumptive use of water or its diversion out of the catchment, and if the focus is directed at the broad seasonal characteristics of the regulated flow regime and not at the detailed consequences of operational policy, then it is reasonably straightforward to describe a simple quantitative measure of the potential degree of flow modification.

It is anticipated, from the information that is available, that the development of reservoir storage for hydro power generation on the Lancang Jiang (Upper Mekong) in Yunnan, China, fits into this simpler non-consumptive category of water resource development and that the major hydrological impact will be a decrease in the characteristic amplitude of the seasonal cycle of flow in the lower basin. The Lower Mekong system is defined as that portion of the total catchment which drains parts of Lao PDR, Myanmar, Thailand, Cambodia and Viet Nam.

This study sets out to assess what these hydrological impacts might be. The only working assumptions are:

• The planned hydro power cascade, involving a reported 23km3 of active reservoir storage on the Mekong in China, will induce significant regulation in the mainstream as it enters the lower basin.

• In terms of its principal features, sensible hypotheses can be advanced as to what the future pattern of these regulated flows might look like at the upstream boundary of the lower basin.

The modified regimes can then be routed downstream using a simple estimate of average monthly tributary inflow between strategic sites on the mainstream.

The Mekong

Depending on the ranking criterion (mean annual flow, catchment area, length of mainstream) the Mekong is around the tenth largest of the world’s rivers. Although a total of over 65M people live in the basin it is, if taken as a whole, sparsely populated. Thirty percent of the total live in the delta region of Viet Nam and a further 15% on the mainstream flood plain and the Tonle Sap regions of Cambodia. That is, almost 45% of the total population live on 27% of the basin land area downstream of the Lao-Cambodian border. This delta/floodplain component receives over 95% of its available water resources from upstream and so there is a natural concern with regard to the consequences of upstream resource development.

With the exception of the right bank tributaries in Thailand, the development of large scale, water resources infrastructure has been slow but the process is accelerating rapidly. The further development of reservoir storage on the Mun and Chi rivers in Thailand, the largest single tributary complex, is unlikely on any significant scale – given that half of the available resource has already been developed and there is a lack of suitable sites. However, there are proposals to meet the rapidly growing Thai urban and industrial demand by diverting water from the Mekong mainstream into the headwaters of the Nam Chi and on into the Nam Mun.

Until recently the only significant reservoir storage on the left bank tributaries was Nam Ngum near Vientiane in Lao PDR, which was completed in 1972 with an active storage of 4.7km3. Since 1996, and the commissioning of the Houey Ho, Theun-Hinboun, Xe Set and Nam Leuk hydro power projects, overall regulation storage in the Lao tributaries has increased by less than half a cubic kilometre, the pondage of these schemes being relatively small. Further hydro power storage is either under construction or planned in the upper Se San in Viet Nam and feasibility studies have been undertaken at sites on the Sre Pok in Cambodia.

Only a small proportion of potential schemes, for which feasibility studies have been completed, have regulation storage in excess of 2km3 and as simulations have shown, the levels of tributary development required for there to be a significant impact on the regime of the mainstream Mekong would have to be impractically high. For example, the hydrological modelling of a scenario in which virtually all proposed tributary hydro power schemes in the Nam Theun, Se Kong and Se San basins (with reservoir storage) are operational, suggests that dry season flows at Kratie (considered here to be the downstream hydrological boundary of the lower system) would increase by 10%. Such a scenario is of course economically naïve and serves only to explore the upper boundary.

Lancang Jiang cascade

In the upper Mekong basin in Yunnan, where the river is referred to as the Lancang Jiang, the water resource focus reverts from the tributaries to the mainstream and the development of a cascade of hydro power regulation on a magnitude larger than even the total existing or proposed storage in the lower system. Details, so far as they are available, are summarised in the table above and indicate that beyond 2020 about 23km3 of regulation storage is planned. The location of this proposed cascade in the context of the wider geography of the Mekong basin is presented opposite.

Some figures have been advanced with regard to the downstream hydrological consequences of such a level of development. For example, Mogg implies that by 2020 the Yunnan cascade will result in a 25% increase in dry season flow and a corresponding 25% decrease in wet season flow on the mainstream in Cambodia and Viet Nam (though the arithmetic behind this statement is not entirely clear). Chapman and Daming report that when Xiaowan is completed before 2010 the mean dry season flow at the Yunnan-Lao PDR border will increase by 40% and by 170% when Nuoxhadu is added to the system. The degree of regulation when Nuoxhadu is commissioned is forecast to be of the order of 20%, that is a 20% transfer of flow volume from the wet to the dry season.

Extensive data on the cascade, provided by the State Power Corporation of China, are quoted in the recent study by Plinston and He Daming (2000). These reveal a total installed capacity of 15,000MW, yielding over 7000MW of firm power and an annual energy output of 70,000GWh.

Hydrological context

For the purposes of hydrological definition, a convenient hydrometric boundary between the upper and the lower basins is provided by the mainstream river gauge at Chiang Saen. This point has the further advantage of being more or less consistent with the wider geographical and political distinction between the two Mekong sub-systems. An equally convenient downstream boundary can be set at the Kratie gauge site, which corresponds to a point where mainstream fluvial behaviour begins to be dominated by hydraulics as opposed to hydrology. At Kratie, 96% of total system flow has already entered the mainstream and the seasonal cycle of water level changes is determined by hydrodynamic behaviour, such as the annual reversal of flows into and out of the Tonle Sap (Great Lake) in Cambodia.

The spatial distribution of runoff in the lower system is summarised in figure 2 which indicates the dominant contribution to mean annual flow in the mainstream of the left bank tributaries in Lao PDR and the Se Kong and Se San rivers, which enter between Pakse and Kratie. Figure 3 presents the incremental growth in mainstream runoff as these major left bank tributaries enter the system taken as a whole. The mean annual flow entering the lower Mekong from China is the equivalent of a relatively modest 450mm depth of runoff. Downstream of Vientiane this increases to 650mm as the principal left bank tributaries, such as the Nam Ngum and Nam Theun, enter the mainstream, only to fall again in response to the right bank entry of the major Nam Chi/Mun system from Thailand. Although this tributary complex drains 20% of the lower system, average annual runoff is only 250mm. Runoff in the mainstream increases again with the entry from the left bank of the Se Kong from southern Lao PDR and Se San and Sre Pok complex from Viet Nam and Cambodia.

In order to evaluate the potential impacts of the Yunnan regulation in the context of this hydrological background, seven strategic sites were selected on the mainstream, downstream of the Lao-China border (figures 1-2). These have 420 station years of hydrological data between them, from which the summary of annual and monthly statistics are given in the tables on pp19-20.

Even such basic data reveal a key feature of Mekong hydrology and one that will serve to amplify the impact of regulation in Yunnan, specifically on dry season flow. As figure 4 shows, the contribution of the Yunnan component (flows entering the system at Chiang Saen) to the hydrology of the lower system is much higher in the dry than in the wet season. A disproportionate volume of the dry season base flow is generated in Yunnan such that, as far downstream as Kratie, it constitutes over 40% of the flow in April. Conversely, in the wet season the proportion falls to below 15%. The implications of this are:

• Large scale reservoir regulation in Yunnan will have a very significant impact on the low flow regime of the lower system mainstream.

• As most of the lower system flow is generated within Lao PDR and Thailand during the wet season, this means that the refilling of large storages upstream of Chaing Saen will not result in such a large reduction in downstream flows, at least once the major left bank tributaries have entered the mainstream in Lao PDR.

Study objective and method

The analysis reported here takes the form of a simple experiment centred on the routing of flow regimes modified by the proposed Yunnan cascade down the mainstream Mekong. The regulated mean seasonal flows are initially defined at the upstream hydrometric boundary of the lower basin at Chiang Saen and are obtained by shifting a proportion of the average wet season flow to the dry season at this point in the system. The magnitude of this seasonal shift is defined here as the degree of regulation.

The long term wet season flow volume at Chiang Saen (1960-1991) is 68km3. Therefore a reallocation of 6.8km3 and 13.6km3 respectively from the wet to the dry season would represent 10% and 20% regulation. These volumes can be assumed to be equivalent to two scenarios of total active storage that have to be refilled at the start of an average wet season. The effect upon the mean monthly flows at Chiang Saen is shown graphically in figure 5 and the modified discharges listed on p20. The graph and table may be interpreted as follows:

•The volumes transferred to the dry season have been added to the unregulated November to April flows in such a way that not only is mean monthly discharge increased but it is also constant over the six months. This aims to replicate the controlled release of water from storage for generating energy.

•This fixed discharge is also assumed to be the controlled hydro power release during the wet season as well. As the storage refills there will be a delay in the rise in the seasonal flood hydrograph which will commence when the storage is full and spillage out of the reservoir system begins. For 10% regulation the start of the annual flood hydrograph at Chiang Saen is delayed by a month, though the timing of average monthly peak discharge, usually August, remains unchanged. For 20% regulation the start of the annual flood is delayed by over two months, the mean annual peak flow reduced by 18% and the timing of this reduced peak set back to September.

Although simplistic, this working definition of regulation is proposed as a transparent and realistic measure of the possible macro impacts of the Yunnan cascade on the seasonal regime of the lower Mekong at its upstream boundary with the Lancang Jiang. The strategy employed to route these impacts downstream is equally simple. The mean monthly lateral inflows to the Lower Mekong for each of the six reaches in between the seven gauged sites indicated on figure 1 are obtained by subtracting the average monthly flow at the upstream site from that at the downstream site. For example, the difference between the mean (unregulated) monthly flows at Chiang Saen and Luang Prabang gives the intervening total tributary inflow. If this is added separately to the regulated monthly flows at Chiang Saen then it will provide an estimate of the potential modification of the flow regime further downstream at Luang Prabang. This process can be continued downstream, reach by reach. The result is a sequence of modified mean seasonal hydrographs indicating the decay of the regulation impact as more and more tributary flows enter the mainstream system.

Results and interpretation

Vientiane in Lao PDR, and the associated right bank towns of Nong Khai and Si Chiang Mai in Thailand, represent one of the largest urban concentrations on the Mekong mainstream upstream of Kratie. Although the catchment area here is some 60% larger than that at Chiang Saen, the major left bank tributaries, such as the Nam Theun, Se Kong and Se San, have yet to enter the main river. The potential impacts of major upstream regulation are therefore considerable, since the hydrology at this point is dominated by the Yunnan component.

As indicated in figure 6 , towards the end of the dry season, low flow would almost double from the natural discharge of 1200m3/sec should regulation in Yunnan involve a 20% reallocation of flow from the wet to the dry season. Even given the lower figure of 10%, flows at the end of the dry season still indicate an increase of 45%. While the upstream storages are refilling during the early wet season, flow in July could fall by as much as 40% and the month of peak discharge move back from August to September. At Kratie, 96% of total flow has already entered the system and yet the dominance of the Yunnan hydrological component with regard to the low flow regime remains much in evidence.

A 20% regulation in China would still result in a 50% increase in average discharge in March, even here, so far downstream. However, by this point on the mainstream the potential impacts on the wet season hydrology are modest.

Figure 7 presents the broader picture and indicates the estimated change in average wet and dry season discharge and how the impacts of regulation might progressively decay downstream. Other than once again underscoring the far greater potential modification of the dry season hydrology, the result also indicates that downstream of Mukdahan, the rate at which the impacts of upstream regulation are progressively masked by lateral inflow to the mainstream, decreases significantly. This is explained by lower average rainfall and runoff in these more southern areas of the basin. This in turn means that any impacts on the flow regime that are evident at Mukdahan, whether the result of regulation in Yunnan or for example, large scale abstraction from the mainstream, would only diminish slowly from this point and onwards downstream.

Comment and conclusion

This simple and exploratory study illustrates that the construction of large scale regulation storage on the Lancang Jiang in Yunnan will have a major impact on the hydrological regime of the entire lower Mekong mainstream. Despite the fact that only 15-20% of the total annual flow that reaches Viet Nam is generated in China, the importance of this Yunnan contribution to the hydrology of the Mekong varies seasonally. It forms the major proportion of dry season flow along most of the mainstream in Lao PDR and Thailand and as far downstream as Cambodia it accounts for almost 45% of the average flow in April. As the wet season progresses from May onwards its contribution to the hydrology of the lower system diminishes significantly as inflows from the major tributaries in Lao PDR and Thailand increase.

On the scale that is proposed, the mainstream regulation in Yunnan has the potential to result in a seasonal reallocation of flow to an extent that average discharge entering the lower Mekong during the dry season could more than double during some months. Such a considerable modification of the low flow regime of the Lancang Jiang (Upper Mekong), combined with the fact that the Yunnan component is the dominant element of the low flow hydrology throughout virtually the entire system has considerable implications.

For example, average dry season flow as far downstream as Cambodia and Viet Nam could increase by up to 20% across the season as a whole, and by up to 50% during March and April. In the middle reaches in Lao PDR and Thailand the increases are obviously even greater.

The potential consequences for flows during the wet season are proportionally less dramatic but still significant. The impacts are likely to be four-fold, namely a reduction in average flood season flow, a decrease in seasonal peak discharge, a delay in the start of the seasonal increase in discharge and potentially the later timing of the peak in an average year. In contrast to the situation in the dry season these impacts are moderated significantly as the relative contribution of the Yunnan component to flood flows diminishes downstream. In Cambodia mean discharge between May and October is estimated to decrease by around 5%, though upstream of Mukdahan in Thailand the changes in flood regime are potentially considerable.

The magnitude of these hydrological impacts will depend on whether or not the entire cascade of reservoirs is actually constructed. Some doubts have been raised about the real prospects for exporting generated power to Thailand, which increases their financial attraction. The key is whether either or both Xiaowan and Nuoxhadu are built. Their reservoirs have proposed regulation storages of 9.8km3 and 12.4km3 respectively, which means that each one can hold the equivalent of over 15% of the average wet season flow at Chiang Saen, close to where the Mekong flows out of China. This in turn means that in principle one or the other has the storage potential to seasonally reallocate 15% of the total flood season flow and both in series to reallocate a massive 22%.

There are many environmental ramifications and impacts to be considered in the development of large dams and reservoirs. Here the focus is exclusively hydrological and in this context it is the potential of the Yunnan cascade to dramatically increase dry season flow that is identified as the major consequence for the regime of the Mekong mainstream as a whole.

Benefits that would result from higher flows between Novem-ber and April include improved navigation along critical reaches, increased flow to the delta region of Viet Nam and the reduced incidence of saline intrusion. Generally there would be an improvement in resource reliability throughout the system during low flow months.