Solar energy and regional coordination as a feasible alternative to the ‘Battery of Asia’ plan

Strategic dam planning and the deployment of decentralized renewable technologies 1 are two elements of the same problem, yet they are normally addressed in isolation. 2 Here, we show that an integrated view of the power system capacity expansion prob3 lem could have transformative effects for the ‘Battery of Asia’ plan. We demon4 strate that Thailand, Laos, and Cambodia have tangible opportunities for meeting 5 projected electricity demand and CO2 emission targets with less hydropower than 6 currently planned—options range from halting the construction of all dams in the 7 Lower Mekong to building 82% of the planned ones. The key enabling strategies 8 for these options to succeed are solar PV and regional coordination, expressed in 9 the form of centralized planning and cross-border power trading. The alternative 10

Here, we introduce a modeling framework for dam and power system planning in 62 the Lower Mekong River Basin that brings the aforementioned elements under the same 63 umbrella. Our framework consists of two components, urbs 29 and VIC-Res 30, 31 . urbs co-64 optimizes capacity expansion for generation, transmission, and storage as well as hourly spatially-distributed hydrologic-hydraulic model simulating not only the relationship be-75 tween hydro-meteorological forcings and water availability thorough the basin, but also 76 the storage dynamics and turbine release of each reservoir. VIC-Res is also implemented 77 for the Chao Phraya, the second main basin of our study site, and home to a few large 78 dams feeding the Lower Mekong power grid. 79 By running our framework over the period 2016-2037, we show that the the regional 80 electricity demand and CO 2 emission targets can be met by constructing only 82% of the 81 planned dams in Thailand, Laos, and Cambodia. The key enabling technologies for this 82 alternative to succeed are solar PV and high-voltage transmission lines, which redistribute 83 cheap electricity across distant load centers. Our analysis of alternative dam portfolios 84 proposes other, more sustainable, options than the 'Battery of Asia' plan 32 : a careful ex-85 pansion of the power system could even absorb the halting of the construction of all dams 86 in the Lower Mekong-at a cost of about 10 billion US$ over the period 2016-2037. 87 Finally, we show that the alternative dam portfolios could substantially limit the fragmen-  countries, reflect a combination of these pathways (Figure 2). In the short term, due to the 104 stringent assumption on the overall emission intensity, we observe that gas replaces part of 105 the coal generation in Thailand and reduces its dependence on imports from Laos. Gas is 106 the cost-efficient solution because the hydropower dams going into operation in 2020 are 107 not sufficient to reduce the carbon emissions in accordance with the stringent targets for 108 that year, and because the installation cost of solar PV is still relatively high (see Table S1 109 for an overview of the technology cost assumptions).

110
The decarbonization strategy shifts drastically from 2025 onwards. The usage of 111 coal is on par with 2016-levels, the relative share of gas decreases, while a huge expansion 112 of renewable energy technologies takes place in the Lower Mekong countries. Since the 113 wind potential is rather limited in the region, the three countries increase the capacities 114 of solar PV (particularly in Thailand) and hydropower (mostly in Laos and Cambodia). 115 Solar PV capacity expansion amounts to 52 GW in 2025 and continues to grow steadily 116 in the following years to reach 68.2 GW by 2037. Thailand alone witnesses an addition of 117 49.8 GW of solar capacity, which is equivalent to about 42% of its total capacity in 2037. 118 Meanwhile, the hydropower capacity in the three countries increases from 9.3 GW in 2016 119 to 22.8 GW in 2037. Most of the new capacities are added in Laos (+12 GW), followed 120 by Cambodia (+1.8 GW), with an additional 0.7 GW in Myanmar dedicated to the Thai 121 power market. This corresponds to an execution rate of 82%, since the total capacity of 122 all planned dams in the region amounts to 17.6 GW. In order to connect the hydropower 123 dams with the demand centers, which are mainly located in Thailand, the power grid is with bioenergy and onshore wind playing minor roles. in the north west of Thailand, whereas the rest is distributed all over the region. How-135 ever, most of the power demand occurs around Bangkok. Hence, we observe that different 136 provinces within the three countries play different roles-notably as hydropower genera-137 tion hubs, solar PV generation hubs, or power demand hubs (the regional distributions of 138 hydro capacities, solar capacities, demand and transmission lines are shown in Figure S1).

139
In order to alleviate the regional discrepancies between supply and demand, the model 140 expands the transmission grid in the east-south direction (from Laos to Thailand through 141 8 Cambodia) and west-south direction (within Thailand), so that most new lines converge 142 towards Bangkok and its surroundings. This cost-optimal power system design implies a 143 high level of regional coordination between the grid operators of the three countries.

144
Impact of alternative dam portfolios on power system expansion. Our results indicate 145 that not all planned hydropower dams must necessarily be built. Moreover, the availabil-146 ity of a vast solar PV potential 15 suggests that there might be opportunities for further 147 reducing the number of dams built in the near future. We therefore consider three alterna-148 tive dam portfolios (Table 1) have a larger impact on migratory fish populations and sediment supply 20, 33 .

154
As illustrated in Figure 3, the alternative dam portfolios are technically feasible, 155 meaning that a decrease in hydropower production can be offset by other sources, mainly 156 solar PV and gas. Interestingly, there is also a positive correlation between hydropower and 157 coal generation. In fact, if the hydropower share is high, then the overall carbon-neutral 158 generation is also high. This leaves some freedom to use coal, which is cheaper than gas 159 with less hydropower, the power system has to generate more energy from carbon-emitting 161 technologies without violating the total CO 2 constraint, so it resorts to using more gas-fired 162 power plants. Importantly, the alternative portfolios may also be economically feasible: 163 Taking into account the investment costs, fuel costs, and fix and variable operation and 164

257
The key message of this paper is that the planned hydropower expansion should be inevitably constrain the number of physical processes and scenarios that can be considered. 286 In other words, screening models deployed across large domains should be complemented 287 by local-scale impact assessments that evaluate additional, fundamental, processes, such 288 as sediment and fish passage through dams 46 . In this regard, another potential modeling 289 avenue is to dynamically link strategic dam planning models and power system planning 290 models, so as to provide a more exhaustive exploration of the ecology-energy trade-offs 44 . 291 Looking forward, it is not difficult to imagine that many developing regions will 292 be caught increasingly in the tension between ensuring cheap power security, exploiting 293 locally available resources, and protecting ecosystems. Multi-model frameworks that span 294 across multiple sectors-like the one described here-are a suitable platform for capturing 295 these multiple perspectives and resolving, or at least addressing, ecology-energy trade-  Table S2.

340
To calibrate the hydrological model, we tuned the parameters controlling the rainfall-  Table S3. in line with the objectives of this study.

422
As explained in the previous section, the hydropower profiles are obtained from VIC-423 Res. They are derived using a single (representative) year. In a sensitivity analysis, we 424 tested the impact of dry and wet conditions on the capacity expansion plans. In Figure S6, 425 we show that the plans are marginally affected by the hydro-climatic variability affecting 426 the region ( Figure S3-S4).

427
No system expansion is allowed in 2016, which is used only for calibration and 428 validation (see Figure S7). We compare the model performance against the projections 429 of the Power Development Plan of Thailand 2018-2037 56 (see Figure S8). We observe 430 minor differences that we are able to explain, and we conclude that the deviations do not 431 affect the main conclusions of this paper. tructure. The RFI is defined as follows 35 :

436
where n is the number of fragments (i.e., river network sections disconnected by dams), 437 v i the volume of the i-th fragment, and V the total river volume (for the entire network).

438
The RFI of a pristine river is 0%, while the one of a totally-disconnected river is 100%.

439
The impact of an individual dam depends on its location as well as the location of other the DOR is defined as 33 : reach with its corresponding river volume, and then aggregating the results for the entire 461 basin 33 : where n is the total number of reaches, DOR i the DOR value of the i-th reach, v i the 464 corresponding volume, and V the total river volume. Note that for a basin affected by 465 multi-year, or carryover, reservoirs, the RRI value can be larger than 100%.