Climate Change Impact - Part 4 - River Mekong (China, Thailand, Myanmar, Lao DPR, Vietnam)

Climate Change Impact

Part 4: Example – River Mekong


A component of a study of the impact of climate change in Cambodia examined how flows in the Mekong River will change in the future. Climate data on precipitation, temperature and other climate variables were used as input to a hydrological model, HYSIM, of the Mekong Basin. The model was calibrated to observed flows at six gauging stations on the main river. The calibrated hydrological model was then used with climate projections to estimate future flows in the River Mekong.


The Mekong River Basin has a drainage area of 795,000 km2and the river is 4350 km in length. The river rises in China at an elevation 5224 m. The river, or its tributaries, also flow through Myanmar, Laos, Thailand, Cambodia and Vietnam. The river’s flow is highly seasonal, dictated by snow melt in the upper reaches and by the Monsoon in the middle and lower reaches.

(At the point where the flow was simulated, Kompong Cham, the basin area is 660,000 km2. It is interesting to compare this with the smallest basin in this series simulated by HYSIM, Pago Stream in US Samoa, which is only 1.52 km2.)

The primary objective of the project was to estimate the impact of climate change on flooding of rural communities and of rural roads in Cambodia. As part of this study a hydrological model of the Mekong River at a daily time step was developed.  The study of the Mekong was necessary for two reasons. Firstly, for communities bordering the river and secondly due its interaction with the inland lake of Tonle Sap.

Figure 1 Mekong River Basin

Current climate

There is a wide variation in the climate over the basin. In particular the temperature in the headwaters of the river are much lower than those closer to the mouth of the river. At Phnom Penh in Cambodia the annual average temperature is 27.4 °C and the range of monthly temperatures is 4.5 °C. At the headwaters of the Mekong the equivalent values are an average of -4.8 °C and a range of 22.7 °C.

Figure 2 Monthly average temperature - Mekong River Basin

To estimate the impact of climate change it was decided to simulate the flows of the Mekong using the HYSIM rainfall runoff model. This model simulates the hydrological and hydraulic process in a river basin with a high degree of physical realism. The model can operate at a daily or shorter time step and in this case, it was decided to simulate the flows at a daily time step. Given the very low temperatures in the upper basin the fact that HYSIM can simulate snow accretion and snow melt was important. The input data required are daily rainfall and daily or monthly potential evapotranspiration (PET). The calculation of PET in turn requires data on temperature, humidity, solar radiation and wind speed.

Thee data came from a variety of sources including:
·         The Ministry of Water Resources and Meteorology of Cambodia - MOWRAM (Flow and climate for Cambodia.)
·         The National Climatic Data Center of the USA. (Daily precipitation and temperature for the whole basin.)
·         The TuTiempo web site (Daily precipitation, temperature, wind speed and relative humidity for the whole basin.)
·         The Climate Research Unit of the University of East Anglia (average monthly values of temperature, relative humidity, wind speed and solar radiation on a 10-minute grid for the whole basin.)

HYSIM has a number of built in data processing apps, these included double-mass plots, infilling of gaps in the data series and the calculation of PET.

Flow data for Cambodia came from MOWRAM and for the rest of the basin from the Global Data Runoff Centre (GRDC).


The first flow measuring station for which data were available was for Chiang Saen, in Thailand immediately downstream of the border with China. The total catchment area at this point is 186,000 km2. However, given the large difference in climate in this part of the basin the catchment was divided in three sub-catchments, each with its own climate data. The following chart shows the simulated daily flow for the period 1988 to 1993 (1993 being the last year with flow data from GRDC site.)

Figure 3 Simulated and observed flow at Chiang Saen

As can be seen, with the exception of 1992 when simulated flows were too high, the simulation is generally accurate.

The simulation was continued downstream with intermediate calibration points at Chiang Khan (Thailand), Mukdahan (Thailand), Pakse (Laos), Stung Treng (Cambodia), Kratie (Cambodia) and Kampong Cham (Cambodia). The following chart shows the flow simulation at Kampong Cham.

Figure 4 Simulated and observed flow at Kampong Cham

For the site in Cambodia data had been ordered for 3 calendar years, 2011 being the last. As can be seen the simulation is generally accurate. There are, evidently, some small differences but given the limited data availability the simulation can be considered satisfactory. There is no doubt that had more time and data been available the simulation could have been improved, in particular if major tributaries had been simulated.

It should also be recognised that the aim of the exercise was to estimate the impact of climate change and the difference in flows.

Climate change

At the time when this study was carried out the latest climate projections based on Representative Concentration Pathways were not available. The earlier SRES projections were used. In this case, based on earlier work in Cambodia, the ECHAM05 model with the A1B projection was used. This option was chosen as the A1B scenario is considered to be the ‘business as usual’ scenario which, given the absence of a successor to the Kyoto protocol limiting CO2 emissions, was appropriate.

The precipitation and temperature were adjusted using projected values of these two parameters. A second form of projection was included based on the work of O’Gorman (Sensitivity of tropical precipitation extremes to climate change. Geophysical Research Letters, published online: 16 September 2012). The paper quantifies the increase in intense precipitation associated with an increase in temperature in the tropics. To use this relationship the daily precipitation values for each calendar year were ranked and the highest precipitation was increased by 10% for each degree of temperature increase and the next two by 6%.

The following chart shows the change in average monthly flow for the River Mekong at Stung Treng, the most upstream flow station in Cambodia.

Figure 5 Projection change in monthly average flow - River Mekong at Stung Treng

This chart shows that, on average, flows in the Mekong will increase as a result of climate change. In particular the flood peak will be higher.

The final chart shows daily simulation of observed flow for the year with average flow, 1987, and projected flows adjusted to represent the increases expected in 2050.

Figure 6 Projected (2050) and observed (1987) flow - Stung Treng


The Mekong is one of the major rivers of the world. This study showed it was possible to accurately simulate flows using data largely in the public domain. When the hydrological model was used with climate projections it was also possible to estimate future flows in the river.


Climate Change Impact - Part 3 - Tonle Sap Lake - Cambodia

Climate Change Impact

Part 3: Example – Tonle Sap Lake - Cambodia


This study related to the estimation of vulnerability (as a function of ‘importance’ and ‘risk’) to climate change of roads and communities surrounding Tonle Sap lake.

A mathematical model of Tonle Sap lake and the channel linking it the River Mekong was developed. This model was able to accurately simulate the lake levels and hence the extent of flooding around the lake; this defined the ‘importance’. Another component of the study estimated the change in levels of the Mekong due to climate change; this identified the ‘risk’. The study drew on the simulation of the Mekong river described elsewhere.

The conclusion was that events which had a rare frequency of occurrence in the past would occur more frequently in the future.


Tonle Sap is the largest lake in South-East Asia, is a wetland of international importance and is recognised by the Ramsar convention. Like most wetlands its area varies significantly through the year, from 2000 km2at its lowest to ten times that figure.  The bed of lake is close to sea level and its maximum level is normally only 10 m above sea level. The distance from the lake to the sea is more than 400 km. The channel from the lake to the Mekong can flow in either direction. When levels in the lake are higher than those in the Mekong water flows out of the lake toward the Mekong (generally from October to April) and for the rest of the year it flows in the opposite direction.

The following map shows three significant locations for level and/or flow measurement. Levels in the lake are recorded at Kampong Loung.  Flow and level in the River Mekong are measured at Kampong Cham and in the Tonle Sap channel are measured at Prek Kdem.

Figure 1- Cambodia and Tonle Sap

Figure 1 - Tonle Sap Lake and Cambodia

Current climate

The fluctuation of levels in Tonle Sap is very much influenced by levels in the Mekong. The levels in the Mekong vary by around 15 m and in the lake by around 7 m. The following chart shows daily water levels in the Mekong at Kampong Cham and in the lake at Kampong Loung. There is approximate synchronicity in the timing of the two sets of levels but with peaks in the Mekong generally being a bit earlier than those in Tonle Sap. This shows that levels in the lake are driven levels in the Mekong.

Figure 2 - Water levels in Tonle Sap Lake and the Mekong at Kampong Cham

A model was then developed which used the levels in the Mekong and the local inflow to the lake to simulated the levels in Tonle Sap.

The flow via the Tonle Sap channel was based on the following equation:

Flow = a * (Mekong level – Tonle Sap level – b)c

If the flows were toward to the lake then this formula was used as above. If it was toward the Mekong then it adjusted by a further factor d.

The values of the four parameters a, b, c and d were obtained by using the ‘Solver’ add-in of Excel. ‘Solver’ adjusts each of the four parameters to see how they change the accuracy of the model. In this case the accuracy of the model is defined as the sum of the squares of the errors in the estimation of the flows in Tonle Sap channel.

The outcome of Solver optimisation process is that the formula became:

Flow = 1126 * (Mekong level – Tonle Sap level – 3.97)1.18

The value of ‘d’, relating to the direction of flow, was 0.64. In reality, this parameter is compensating for some hydraulic factors not included in this model. A full solution of the equations would take account of the inertia of the water in the Tonle Sap channel; in simple terms when the relative levels in the lake and the Mekong change they first have to stop the river flowing in one direction before they can increase its flow in the opposite direction.

The value of parameter ‘b’, 3.97 m, which allows for the difference in the datum at Kampong Cham and at Prek Kdam is compatible with the figure of water levels above.

The following chart shows the simulated and observed water level in Tonle Sap.

Figure 3- Simulated and observed levels in Tonle Sap Lake

As can be seen the simulation is generally accurate. Many of the peaks of water level are slightly underestimated but otherwise it is good. The correlation coefficient between observed and simulated levels is 0.967.

It can therefore be concluded that the simulation of water levels in Tonle Sap Lake is sufficiently accurate for the model of lake levels to be used to study flooding around the lake.


The aim of vulnerability mapping is to identify locations at risk where interventions to reduce vulnerability are needed. Two factors are involved. The first is the importance of the risk; if a road connects large communities, for example, it is more important that it continues to function and serve a wider community than a road with lower importance.

A simple definition of vulnerability is:

Vulnerability = Importance x Risk  …………………… Equation 1

Importance of road segments

To evaluate the importance of a road, a scoring system was developed.  The aim was to be able to identify the importance of a road. It is appreciated that such an ideal will never be completely achieved; whatever the algorithm says each would have to be examined using the calculated value as a guide. The scoring was applied to each road section, defined as a section of road between two junctions. In all there were 5263 road sections. They covered 8 provinces.

The Ministry of Rural Development of Cambodia (MRD) already has a system of road classification which goes from 1, the most important, to 4, the least important. As the aim was to have a higher weighting for higher importance this number system was reversed.

Another factor is what the road connects to. A road with a low ranking could be considered more important if it joined a road of higher rank.

The length of a road is also a factor – the longer the road the more important it could be considered. 
To have a scoring system compatible with the numbers associated with road category, the logarithm of the length in metres was used. This would go from 2 for a road of 100 metres up 4 for a road of 10,000 metres.

The population served by the road is also a significant factor. As the data on communes identifies the area and the coordinates of the centre, the algorithm identified communes based on the square root of the area (approximately the distance from the centre to an edge of the commune and distance from the road. To have a score compatible with other elements the logarithm of the population was used as a score. In this case the population was the total of all communes adjoining the road.

The data base also lists wats (pagodas), mosques and churches. Since wats are usually built on high ground and provide refuge during a flood the presence of a wat was given a score of 3.

Health centres, which are important to the whole community but not specifically relate to flooding were given a score of 2.

The final element was the presence of a school for which the score was 1.

The following table summarises the scoring system

Table 1 - Summary of road importance scoring
Road segment identification
Coordinates and brief description
For cross-referencing only.
Road category
MRD categories from 1 to 4
Score in inverse order. Class 1 has 4 points, class 4 has 1 point.
Logarithm of the road length in metres. For example, a segment 10,000 metres long world have a score of 4
Category of road joined to
e.g. National Road, MRD 3
As for road categories. If connected to a national road then 5 points. Points are given for both ends.
Population in communes adjoining road
This is the total population of all the communes the road passes through.
Score is based on the logarithm of the population. For example, if the population is 30,000 the score is 4.5
The presence or otherwise of a school
Score is 1 or 0
The presence or otherwise of a Wat (Pagoda)
Score is 3 or 0. Higher than a school as it relates to the whole community and often provides a refuge during a flood.
Clinic or health centre
The presence or otherwise of a clinic.
Score is 2 or 0. Higher than a school as it relates to the whole community.

The scoring system was applied to MRD roads in eight provinces: Battambong, Kampong Cham, Kampong Chhnang, Kampong Speu, Kampong Thom, Pursat,  Siem Reap         and Thboung_Khmom. Tis established the ‘importance’ element of equation 1.

Impact of climate change

The model of Tonle Sap, combined with modelling of flow in the Mekong under current project (climate change) values established the ‘risk’.
In general terms, the following summarise the projected changes in road vulnerability:
  •         What was a 1 in 5 year flood is projected to occur every 2 years.
  •         What was a 1 in 10 year flood is projected to occur every 3 years.
  •         What was a 1 in 25 year flood is projected to occur every 6 years.
  •         What was a 1 in 100 year flood is projected to occur every 9 years.
  •          A 1 in 100 year flood is projected to increase in area by 10%.

See Older Posts...