Map of (A) the large dams (height > 15 m) in the Mekong River Basin and the (B) existing, under construction and planned hydropower projects in the Upper Mekong Basin (UMB). The hydrological stations for discharge analyses are shown with blue triangles (A). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
· * The Upper Mekong Basin is undergoing extensive hydropower development and its largest dams have recently become operational. In this study we assess the discharge changes using observed river discharge data and a distributed hydrological model over the period of 1960–2014. Our findings indicate that the hydropower operations have considerably modified the river discharges since 2011 and the largest changes were observed in 2014. According to observed and simulated discharges, the most notable changes occurred in northern Thailand (Chiang Saen) in March-May 2014 when the discharge increased by 121–187% and in July-August 2014 when the discharge decreased by 32–46% compared to average discharges. The respective changes in Cambodia (Kratie) were 41–74% increase in March-May 2014 and 0–6% decrease in July-August 2014 discharges.
· * The key driver for the ecological productivity of the Mekong River is the annual flood pulse resulting from the seasonal monsoon climate (Holtgrieve et al., 2013; Junk et al., 1989; Lamberts, 2008 ; MRC, 2010). The Mekong’s flood pulse is characterised by distinct low flow season in December-May and a high flow season in June-November (MRC, 2005). The flood pulse sustains base functions of ecological productivity by transporting vast amounts of sediments and nutrients and inundating extensive floodplains, and also by providing a diversity of ecological habitats (Junk et al., 1989 ; Junk et al., 2006), such as the Tonle Sap Lake in Cambodia (Arias et al., 2014; Arias et al., 2012 ; Holtgrieve et al., 2013). However, the ongoing hydropower development is likely to considerably affect the flow regimes and the annual flood pulse. The amplitude of the annual flood pulse is expected to be reduced (Lauri et al., 2012 ; Piman et al., 2013) and thus affect the sediment and nutrient transport, flood extent and related ecological habitats (Arias et al., 2014; Arias et al., 2012 ; Holtgrieve et al., 2013). Moreover, the dams and reservoirs trap large quantities of sediments and nutrients (Kondolf et al., 2014; Kummu et al., 2010 ; Maavara et al., 2015), further decreasing the sediment load after observed falling trend in it due to lower tropical-cyclone activity (Darby et al., 2016). Moreover, dams are reported to severely prevent fish migration in the Mekong (Baran and Myschowoda, 2009 ; Ziv et al., 2012).
· * The largest hydropower projects in the Mekong are built into the Upper Mekong Basin (UMB), where the river is also known as the Lancang River. The UMB has currently seven completed large dams and twenty more under construction or planned (Fig. 1 and Table 1) (see also Hennig et al., 2013). The hydropower development in the UMB started by construction of Manwan dam, which became operational in 1993 and was fully completed in 1995. The most recent development is the completion of Nuozhadu dam in 2014, which is the largest hydropower project in the whole Mekong Basin (Fig. 1 and Table 1). The hydropower development in the UMB is driven by increasing energy demand in the province of Yunnan and in the eastern parts of China and the need to reduce emissions from energy production (Chang et al., 2010; Chen et al., 2010 ; Hennig et al., 2013). The construction of large scale hydropower in the UMB has raised considerable concerns regarding the impacts on the downstream countries (e.g. Kattelus et al., 2015 ; Kuenzer et al., 2013).
* The future impacts of dams in the UMB on dowstream river discharge are estimated at least in three studies (Hoanh et al., 2010; Lauri et al., 2012 ; Räsänen et al., 2012). These studies estimate discharge changes caused by Manwan, Dachaoshan, Gonguoqia, Jinghong, Xiaowan, Nuozhadu hydropower projects (Fig. 1 and Table 1) at Chiang Saen using model-based approaches. These studies estimate that at Chiang Saen, the dry season discharge (Dec-May) is predicted to increase by 60–90% and the wet season discharge (Jun-Nov) to decrease by 17–22% (Hoanh et al., 2010; Lauri et al., 2012 ; Räsänen et al., 2012). These changes are a result of storing of water into reservoirs during the wet season and releases during the dry season. The model-based estimates further show that the UMB hydropower operations are subject to increase (decrease) the dry (wet) season discharge variability, and that the discharge impacts are observable as far downstream as Kratie in Cambodia (see location at Fig. 1) (Räsänen et al., 2012). Piman et al. (2013) also provide estimates of hydrological changes from basin-wide development scenarios, but they do not report changes at Chiang Saen and specifically for hydropower development in the UMB. These model simulations were performed with pre-2011 datasets and thus are not validated against observed river discharge changes. In addition to above analyses, the cumulative impacts of Manwan dam and climate have been studied (Zhao et al., 2013). The Manwan dam is found to compensate for climate variability.
* The discharge data were obtained from Mekong River Commission Secretariat’s quality assured database and the analysis period was limited by the data availability from Chiang Saen station. We used discharge data instead of water level data as the river geomorphology may have changed in time and the rating curves have been updated in several occasions. We did not have information of all rating curves and therefore we analysed the data for possible inconsistencies between different periods with different rating curves and did not find problems. We used the discharge data as such. We chose these three hydrological stations for following reasons: (i) availability of the data is continuous and long enough; (ii) Chiang Saen is the most upstream station in the LMB and it captures well the hydropower operations along the UMB; and (iii) Nakhon Phanom and Kratie show how the discharge anomalies propagate along the Mekong River. Downstream of Kratie the discharge regime becomes smoother and less variable as the discharge condition is dominated by hydraulic features in addition to hydrologic condition.
· * For estimating the impacts of hydropower operations in the UMB on the river discharges at Nakhon Phanom and Kratie, the impacts of the dams in the LMB needed to be excluded. This was done by simulating the hydrology of the whole Mekong Basin and by estimating both natural and regulated discharges at Nakhon Phanom and Kratie. We first simulated the natural (non-regulated) discharge over the period 1998–2014 (scenario C1; see Table 2), which represented situation without any hydropower operations in the whole Mekong Basin. This was done with the distributed hydrological model VMod, similarly as for the UMB and Chiang Saen for scenario B1. Then we included the observed impacts of hydropower operations in the UMB to the simulation by using the observed discharge at Chiang Saen (1998–2014, discharge scenario A1) as an input to the model at that location, resulting to scenario C2. The hydrological model simulation thus allowed us to estimate how the observed impacts of hydropower operations at Chiang Saen propagate downstream to Nakhon Phanom and Kratie, while simulating the natural, climate driven hydrology in the LMB.
· The comparison of simulated natural discharge (Scenario C1) with the simulated discharge with impacts of hydropower operation in the UMB (C2) over the period of 1998–2014 reveals how the discharge anomalies observed at Chiang Saen propagate downstream. During the years of 2013 and 2014 the discharge changes observed at Chiang Saen caused considerable discharge changes at Nakhon Phanom and Kratie (Figs. 4 and 5, and Table 3). The discharge impacts are stronger at upstream station Nakhon Phanom than at Kratie.
· * The general nature of changes in simulated discharge were similar at Nakhon Phanom and Kratie as the observed changes at Chiang Saen: dry season discharges increased and wet season discharges decreased. The largest discharge changes were found at Nakhon Phanom and Kratie in March-May 2014 when the simulations suggest 88–118% and 41–68% increase in monthly average discharges, respectively, compared to simulated natural non-regulated discharge (Table 3). During the wet season, the largest changes were found at Nakhon Phanom and Kratie in July-September 2014 when the monthly average discharges decreased 3–11% and 0–6%, respectively.
* We identified eight record high and low discharge events in the observed discharge at Chiang Saen during the period of 2010–2014, when compared to period of 1960–1990, i.e. before hydropower development in the UMB (Section 3.1). The model simulations allowed us to separate the climate induced impacts from anthropogenic ones (Section 3.1). According to these simulations, the record events were not caused by climate or weather events, but resulted from hydropower operations in the UMB (Fig. 3). The record high discharges occurred in March-April 2011, February-April 2013 and in January-April 2014, while the record low discharges occurred in February-March 2010, June-July 2012, October 2013, July 2014 and in August 2014 (Fig. 2). The basin-wide simulations, in turn, revealed that the identified record high and low discharges propagated downstream the Mekong causing discharge anomalies as far as Kratie in Cambodia (Section 3.3, Table 3). Thus, our results support the literature (e.g. Räsänen et al., 2012) that the hydropower operations in the UMB have caused considerable changes in the discharge regime of the Mekong River.
* In addition, the results suggest that the hydropower operation in UMB are only partially responsible for the observed discharge changes in Kratie. For example, during the dry season of 2014 record high discharge was observed in Kratie (Fig. 2C), but according to our simulations the hydropower operations in UMB could explain only half of the observed discharge increase (Table 3). The LMB has currently 45 large operational dams (CGIAR WLE, 2015 ; MRC, 2015) (Fig. 1) and it is possible that they have contributed to the increase in dry season discharge in Kratie in 2014.
· * The impacts of hydropower operations in the UMB on river discharges at Kratie are estimated earlier only by Räsänen et al. (2012) and the current research is the first one to estimate those on the base of observed discharge data. Our estimates (scenario C) suggest smaller discharge change in February and larger discharge changes in April-June than the estimation by Räsänen et al. (2012) as shown in Fig. 6C. Otherwise, the estimates suggest similar pattern of change. The differences between our estimate and estimate in Räsänen et al. (2012) are likely a result, at least partially, of the comparison method. We compared our estimate of discharge change during one year against the average discharge change of seven years of simulated hydropower operations in Räsänen et al. (2012). Altogether, the comparison at Chiang Saen and Kratie shows that the observed river discharge impacts of hydropower operations in the UMB were as expected, but they suggest that the discharge impacts may vary from year to year, depending on variations in hydropower operations and climate and weather conditions.
· * Hydrological consequences: The discovered discharge impacts of hydropower in the UMB reflect the hydropower operations: the reservoirs store water during the wet season and release it during the dry season thus increase dry season flows and reduce the wet season flows. These changes resulted in dampening of the Mekong’s annual flood pulse, particularly in the upper parts of the LMB. The decreased amplitude of the flood pulse is expected to reduce the sediment and nutrient transport and negatively affect the aquatic habitats that depend on large seasonal water level fluctuation (see e.g. Lamberts, 2008). The increased dry season flows are also expected to change the habitat conditions in river channels through increased flow velocity and water depth and decreased sunlight penetration (Bunn and Arthington, 2002).
The results also show large and rapid water level variations during the dry season, particularly after the completion of the largest dams in the UMB (see e.g. Fig. 2). This is expected to result in increased instability in the river geomorphology (Brandt, 2000) and in aquatic (Bunn and Arthington, 2002) and riparian habitats (Nilsson and Berggren, 2000).
The major concern is that the flood pulse and discharge changes threaten the Mekong’s ecological productivity, which is the basis for livelihood, income and food security for millions people (Friend et al., 2009 ; MRC, 2010). For example, the Mekong River is one of the world’s most productive inland fisheries and fishing is one of the most important livelihoods in the region. In food security context, fish and aquatic animals are in key role as they provide 47–80% of consumed animal protein in the LMB countries (Hortle, 2007). In addition, the changes in annual flood pulse and short-term water level variations affect the agricultural activities in flood plains and riverbanks. The changes in annual flood pulse may also bring new opportunities for the LMB countries. The increased dry season discharge increases the water availability in the dry season, which may attract further hydropower development, provide water for irrigation, and improve navigation possibilities (ICEM, 2010). However, the results also showed that the dry season discharge may vary considerably due to hydropower operations, as predicted also by Räsänen et al. (2012), and this may cause challenges for downstream development. Altogether, the results reinforce the need for developing upstream-downstream cooperation in terms of water resources management and development.
· * Future research directions: Three future research directions are identified.
First, there is a need to monitor and further estimate the future impacts of hydropower development in the UMB. This analysis was based on time period (year 2014) when there were only seven operational hydropower projects in UMB and the largest hydropower projects had only recently become fully operational (Table 1). The total regulating capacity of the seven existing hydropower projects is over 23 km3 corresponding to 27% of the annual discharge at Chiang Saen, 10% at Nakhon Phanom and 6% at Kratie. The future plans include 20 more dams and they are likely to aggravate the already observed downstream impacts as the regulating capacity in the UMB will increase in the future.
Secondly, the variations in climate affect the hydropower operations and their cumulative impacts are less studied. For example, during the dry season of 2010 the Mekong River experienced exceptionally low water levels (Stone, 2010) and our findings suggest that the low water levels, at least partially, resulted from hydropower operations. Räsänen et al. (2012) also suggested that climate variability would affect hydropower operations and together they may increase variability in river discharges, particularly the inter-annual variability of seasonal discharges (see also Wang et al., 2006). The climate change is also expected to affect the hydrology in the Mekong River, but how it will affect the discharge regimes is uncertain (Hoang et al., 2016; Kingston et al., 2011 ; Lauri et al., 2012). The cumulative impacts of climate and hydropower operations are likely to become more important in the future as the water resources are developing in all countries of the Mekong River and the water is projected to become an increasingly competed resource.
Thirdly, the scientific empirical evidence on the effects of river flow alteration on the aquatic ecology has remained limited in the Mekong River. Some research is conducted, for example in the Tonle Sap lake-floodplain system (Arias et al., 2013), but the empirical research in river channels is scarce, not published or non-existing. The empirical research would be highly important as it enables the development of case specific and generalised assessment methodologies for assessing the impacts of future hydropower development and thus provides guidance for developing hydropower into more sustainable direction. For example, even small changes in hydropower operations can be beneficial for aquatic ecosystems (Poff and Schmidt, 2016).
* Our findings indicate that since the year 2011, the hydropower operations in the Upper Mekong Basin have resulted in considerable discharge changes throughout the Mekong River, as far as in Cambodia. General changes were characterised by increase in dry season discharges, and decrease in wet season discharges, and large and rapid discharge fluctuations in the dry season. Largest changes were observed in 2014 after the completion of Nuozhadu dam, the largest hydropower project in the whole Mekong Basin. The earlier model-based predictions on the discharge changes in the scientific literature are in line with the discharge changes presented in this paper, although the observed changes in 2014 are partly larger than the earlier predictions suggest. The findings further show that the hydropower operations in the Upper Mekong Basin can explain only partially observed river discharge changes in Cambodia (Kratie), which suggests that river discharges are affected also by dam operations in the Lower Mekong Basin. The impacts of hydropower development in the UMB on downstream discharge are expected to vary from year to year depending on the water availability and hydropower operations. Further, they are expected to increase in the future as more dams are planned to be built.
* The observed river discharge changes may have major implications to downstream countries and people. The major concern is that the flow changes affect negatively the ecological productivity of the river system and thus affect the water related livelihoods and economic activities in the downstream countries. The upstream hydropower development may also increase downstream water availability, which in turn can provide favourable opportunities for irrigation and make the downstream hydropower development more attractive. However, the utilisation of river flows in downstream countries may face challenges due to unpredictable upstream flow regulations. The future challenge is how to maintain the ecological functions of the river and share the benefits and losses of hydropower development equitably. Addressing the negative impacts is urgent and it requires strong transboundary cooperation.
* Acknowledgements: We would like to thank the Mekong River Commission Secretariat for providing the observed discharge data and the dam database for the Mekong River Basin. We would also like to thank Kim Geheb at WLE Greater Mekong (CGIAR Research Program on Water, Land and Ecosystems) for providing their dam database for the Mekong River Basin. TAR received funding from Maa- ja vesitekniikan tuki ry. and MK from Academy of Finland project SCART (grant no. 267463) and Emil Aaltonen Foundation project ‘eat-less-water’.
ARTICLE ON GOOGLE DRIVE: