Climate Change Impact on the Streamflow for the Nam Ou Catchment, North Laos in the Mekong Basin

Southeast Asia, in general, and the Mekong Basin (MB), in particular, with its typically warm and wet climate, face water resource challenges. A deep understanding of the future streamflow is needed to manage water resource suc - cessfully. Data scarcity and topographical differences have made it difficult to accurately reproduce the streamflow regime in the sub-catchment of the MB. The main goal of this study was to provide the first assessments of stream - flow impacts due to climate change for the Nam Ou Basin, a primary Lao sub-catchment of the MB, employing the most updated Couple Model Intercomparison Project Phase 6 (CMIP6) climate scenarios. The MIKE-NAM (Nedbor Affstromnings Model), the observed hydro-meteorological data, and the Moderate Resolution Imaging Spectroradiometer (MODIS) evaporation were employed. The climate change scenarios showed increases in seasonal and annual river discharges with different magnitudes in the future. The annual streamflow was expected to rise by 0.31%, 16.75%, and 38.31% in the 2040s as well as 23.35%, 32.80%, and 74.82% in the 2080s under three scenarios, respectively. The wet season in the Nam Ou Basin occurs one month earlier. The wet season flows increased by 5.6–76.9%, and the dry season flow showed a contrasting directional change, decreased by 8.4%. The annual peak discharge also exhibited an increase of 3.2–14.6% for the SSP1-1.9 scenario in the mid-century (the 2040s), and end-century (2080s). Those figures are 8.9–19.7% for the SSP2-4.5, and 23.3–48.9% for the SSP5-8.5 scenario, respectively. The study revealed the streamflow variation under the effect of climate change in the Nam Ou Basin, a sub-catchment of the MB, highlighting the need for special consideration in disaster risk mitigation, especially under climate change.


INTRODUCTION
It is projected that evapotranspiration and the pattern of precipitation will change due to global warming. Therefore, they will affect the streamflow regime and increase the challenge for sustainable water resources management (Hu et al., 2013, Lu, 2005. Many studies found that under global warming, the severe weather like drought, heavy precipitation, and the stream-  (Shrestha et al. 2013(Shrestha et al. , 2016, Sesan, Sekong, and Srepok Basins in Vietnam (Oeurng et al. 2016). River flow is anticipated to increase in the wet and reduce in the dry season.
In order to evaluate streamflow variation under climate scenarios, the main variables, namely temperature and precipitation, are used as input for hydrology modeling. This approach is the most common and reliable for climate change impact assessments on hydrology (Li et 2018) found that the sediment discharge in the basin increases under climate change and consequently decreases the reservoir's capacity. However, construction of reservoirs will reverse that shift and lowering the sediment outflow by 44 to 89%. These findings suggested that extreme occurrences (floods) are more frequent and intense. These projections provide insight into the hydrological changes caused by climate change, significantly affecting socioeconomic and ecological development.
All the existing studies used the future climate conditions of previous scenarios. CMIP Phase 6 or CMIP6, which is the most current climate change scenario, showing essential updates to the horizontal resolution and the emission mechanism in the scenarios (Peng et al., 2023, O'Neil et al., 2020, has not yet been used to investigate streamflow impacts due to climate change for the Mekong or it sub-basins. Therefore, this study used an ensemble mean of CMIP6 scenarios to provide critical hydrological knowledge for the Nam Ou sub-catchment of the MB. One of the most basic and reliable tools to address the streamflow response under various climate change scenarios of the above-mentioned hydrologic models is the rainfall-runoff model, MIKE-NAM (Golmohammadi et al., 2014;DHI 2009). Its key advantage is the ability to model quickly with minimal input requirements. This advantage makes it a good option for developing countries with inadequate observations for hydrologic modeling, such as the Nam Ou River basin.
The Mekong River in Southeast Asia's tropical region is especially vulnerable to the consequences of climate change (Shrestha et al. 2016). Undeveloped nations, such as Laos, are even more vulnerable to droughts and floods due to the dramatic climate change in Southeast Asia. The Nam Ou basin is one of the essential Mekong sub-catchments, and is Laos's largest catchment. It supplies approximately 12.2 × 109 m 3 of water to the MB annually, making it the sixth largest river regarding streamflow (Peter-John Meynell 2016). Therefore, variations in the Nam Ou River's flows may cause adverse effects downstream of the Mekong River. Even though recent studies have started investigating these concerns (Shrestha et al. , the river discharge impact due to climate change of the Nam Ou basin is mainly unidentified. Since the severity of severe weather is expected to grow, an updated evaluation of streamflow impacts due to climate is crucial for water resources management in the Nam Ou catchment and the broader MB (Lauri et al. 2012).
This study aimed to provide the first assessment of streamflow impact due to climate change on annual and seasonal scales for the Nam Ou Basin to mitigate the water-related challenges in the research study.

Study area
The Nam Ou River runs over 448 km. It is an essential Mekong River tributary ( Fig. 1). It begins near the Lao-Chinese border in Muang Ou Nua and crosses the northern Laos mountains until the Mekong River in Ban Pak Ou (MRC 2011). The Nam Ou Basin area is 20087 km 2 , with an altitude range of 129-1946 m. The Nam Ou Basin experiences average annual precipitation from 1650 mm to 1950 mm. The monthly average temperature is 27.3°C (between 24°C and 29°C). This area is dominated by the dry season, which lasts from November to April, and the wet (rainy) season, which lasts from May to October, with 90% of precipitation occurring (MRC 2011).

MIKE NAM model
The MIKE-NAM is a lumped-parameter rainfall-runoff model and requires a little input data. The MIKE-NAM model illustrates the runoff generation process by simplifying the catchment to four separate connected reservoirs with representing parameters. Nine default parameters representing surface-root zone and groundwater were calibrated and validated. The initial conditions used to depict the condition of the basin at the start of the rainfall event include the initial value of overland flow, water contents in surface and root zone storage, and baseflow and interflow. It is usually best to run the model when the dry season ends with the initial conditions set to zero, excluding root zone storage and baseflow value. Water content in the root zone storage usually occupies 10-30% of the capacity, and the baseflow value is close to the observed one (DHI 2017). This study did not consider irrigation and snowmelt parameters due to the climate features and unavailable irrigation information.
Model parameters are calibrated against discharge observations using automatic procedures. The correlation coefficient (), and the water balance error () given by the below formulas are used to evaluate the performance of the model: where: N -indicates the sample size; ̅̅̅ ̅̅̅̅the mean discharge of simulated and observed one, respectively; Q s,i , Q o,i -the samples i of simulated and observed discharge. The model works well when the R reaches unity and WB er approaches zero.
After the model's calibration and validation have been implemented, the model with the calibrated parameters is applied to evaluate streamflow response based on climate change scenarios.

Precipitation
There are three rain gauges in the Nam Ou Basin: Phongsaly, Oudomxay, and Muong Ngoy. The monthly precipitation (mm) time series of these three gauges are collected from the Mekong River Commission (MRC) (Fig. 1). The periods from 1998-2000 and 2002-2003 were applied for calibration and validation, respectively. A Thiessen polygon method integrated into Arc GIS 10.5 software generates the weighted rainfall average for a catchment. The resulting percentage weights of those rain gauges are displayed in Table 1.

Evaporation
The MOD16A2GF Version 6 product acquired monthly evaporation data at 500 m resolution. The MOD16 data product is based on daily meteorological reanalysis data, the Penman-Monteith equation, and Moderate Resolution Imaging Spectroradiometer (MODIS) data (i.e., vegetation property dynamics, albedo, and land cover) obtained from the MRC was applied for calibration and validation (Fig. 1).

Climate change scenarios
The future precipitation generated by the precipitation percent change for mid (the 2040s) and far future (2080s) according to three SSPs (1-1.9, 2-4.5, and 5-8.5). The observed precipitation of the baseline 1999-2008 (the 2000s) multiplied with the proportions (i.e., the precipitation percent change) to generate the precipitation for each rain gauge. The mean of the multi-model CMIP6 ensemble for sub-national aggregations was employed (WB 2022). The precipitation percent change using average level is a simple but widely used technique. It is worth noting that selecting a baseline period representing the current climate conditions plays a particular role in researching the impact of climate change. The longer the rainfall time series is, the more representing the climate are. However, based on the maximum data that can be collected for the study

MIKE-NAM performance
A computer-based automatic calibration procedure was used to calibrate the model's parameter using the daily data from 1/1/1999 to 31/12/2000. The validation period was from 1/1/2002 to 31/12/2003. The ten parameters of MIKE-NAM were determined through observed discharge of the Muong Ngoy station ( Table 2).
The model's performance has been evaluated both visually and quantitatively. Figure 2 compares simulated and observed discharge at the Muong Ngoy station for verification. Both calibrated and validated results show that the simulation at the Muong Ngoy station highly resembles its observation. The mathematical equation of evaluation criteria is provided in Table 3. The values were 12.4% (calibration) and 1.6% (validation), indicating that the simulated discharge was typically underestimated compared with the observation. The calibrated and validated were also 0.77 and 0.87, respectively. Regarding calibration (1999)(2000) and validation (2002)(2003), the MIKE-NAM model can generate the discharge with a high matching between the observation and simulation for the entire Nam Ou basin.

Projected precipitation
Under all three Shared Socioeconomic Pathways scenarios (SSPs), the monthly precipitation during the rainy season (from July to October) is predicted to rise consistently in the 2040s and 2080s across all provinces (Fig. 3). Moreover, it is expected that precipitation will decline mainly in January, February, May, June, November, and December in both time frames. The precipitation in February in SSP2-4.5 and SSP5-8.5 will drop during the 2040s and 2080s but increase in all remaining time slices under the three SSPs. The

Streamflow change
Using the calibrated parameters, the discharge was projected in response to climate conditions under SSPs using the MIKE-NAM model. The annual flows increase during 2040-2099 under all three SSPs (Fig. 4). The SSP5-8.5 scenario's climate conditions resulted in the largest increase in streamflow (74.82%) from the 2020s to 2080s when compared to the baseline. The smallest increase was seen in the streamflow under the SSP1-1.9, which went from 472.44 m 3 /s in the baseline to 473.89 m 3 /s under the SSP1-1.9 (only 0.31% increase). Figure 5 displays annual, wet, and dry season flow changes under SSPs. The annual flow increased from 0.31% to 38.31% in the 2040s (Fig. 5a) and increased by 23.35%, 32.80%, and 74.82% in the 2080s under all SSPs. There was a significant difference in streamflow variation between seasons to climate change. Wet season flow increased (Fig. 5b) from 5.56% to 43.30% in the 2040s and from 24.75% to 76.89% in the 2080s   in the future. The dry season flow decreased only in the 2040s under SSP1-1.9 scenarios (Fig. 5c). Figure 5 displays the variation in monthly discharge for all SSPs over the two upcoming decades (the 2040s and 2080s). On the monthly scale, the different responses between scenarios were observed. For the upcoming decades, flow increased from April to July under all SSPs, with the most significant increase in June (Fig. 5). In contrast, the flow was expected to drop in January, February, August, September, October, November, and December in all future decades under SSP1-1.9 -with the largest decrease in September (Fig. 5). This was connected to the reduction in dry precipitation. Notably, the decreases in streamflow under SSP1-1.9 for the 2040s exceeded 20% in all dry months. Compared to the baseline period, streamflow increased in all months for all future decades under SSP5-8.5 and 2080s under SSP2-4.5, except for January projected to The decreased flows from November to March under the three SSPs during the 2040s may be caused by an expected decline in the dry month precipitation. This precipitation decline will probably result in an earlier wet season under climate change. A warmer climate (under the SSP5-8.5 scenario condition), with lower and higher rainfall at different places, results in higher streamflow in all-time slides ( Fig. 5a and  b). SP1-1.9 shows a slight decrease of 3.24% in the mid-future and an increase of 14.6% in the far future. In all evaluated periods, the peak discharge indicated a significant increase. Increases amounting to 8.9% and 19.7% are observed in the annual peak discharge at Muong Ngoy for SSP2-4.5, respectively, at the mid and end century. A sharply increased of 23.3% and 48.9% in the peak discharge is seen for the two future projection periods in the SSP5-8.5.
Since this study is the first evaluation of the streamflow variation using the CMIP6 and the MIKE-NAM model for the Nam Ou basin, its performance is assessed by contrasting its results to the previous related research for the MB. These results align with earlier research on the effects of climate change in the MB. An increase of 23%, 16%, and 36% in streamflow are observed in SSP1-1.9, 2-4.5, and 5-8.5, resulting from a 12%, 7%, and 16% increase in precipitation. In comparison, Li

DISCUSSION
This study has presented the streamflow variations in the Nam Ou River basin based on an ensemble mean of CMIP6 scenarios. As such, these results provide critical updates to our understanding of how streamflow responds to climate change which is vital for adapting water resources under climate change. The streamflow response in the Nam Ou basin is generally consistent with the results of earlier research on climate change which observed increasing trends in monthly and wet season discharges. This analysis expects an increase of +0.31% -+75% in annual discharge at Nam Ou, compared to -17 -+66% by Shrestha et al. (2013). The wet season discharge and the yearly peak value will increase in all scenarios, but the low flow will reduce in the SSP1-1.9 scenario. These worsening trends can potentially increase water risks across the basin and affect the downstream area.
The approach used in this study has some drawbacks. The change factor method is used to project future climate conditions. Only scales of the mean of precipitation variables were used. The spatial pattern and the temporal sequence of wet days are assumed unchanged. However, this approach only requires information at a monthly time scale suitable for the data-scarce region as the Nam Ou Basin.

CONCLUSIONS
This study is the first climate change impact assessment on the streamflow for the Nam Ou river basin, a sub-catchment of MB in Lao PDR, based on an ensemble mean of CMIP6 scenarios. The results provide essential updates to understanding the streamflow variation under climate change for the data-scarce region, the Nam Ou basin. In most cases, precipitation increases by 7.7-15.8%, but certain areas and certain months show declining signals. As a result, all scenarios show an increase in annual river discharge, from 0.31 to 74.82%. Increases in both wet and dry seasons are also depicted under climate change impact. The discharge of the rainy season is more significantly impacted by climate change than the dry season. However, the discharge changes in different direction under the SSP1-1.9 and 2-4.5. The wet stream flow was expected to increase by 27.7-112.9% during 2040-2080. The dry season (from August to March) decreases by 1.06-31.53% in the mid-century under the SSP1-1.9. The wet season tends to occur one month earlier in the future. These changes imply significant consequences for controlling water risk in the downstream river delta.