TL;DR
This paper presents a high-resolution mesoscale simulation of wind power potential across Guangdong Province, China, using the MM5/CALMET system. The study reveals that coastal areas in Guangdong possess significant wind power potential, with wind speeds exceeding 5 m/s for approximately 70% of the year, making it a promising region for wind energy development.
What Problem Does This Paper Address?
Accurate assessment of wind power potential is crucial for the successful planning and implementation of wind energy projects. However, regions with complex topography, like Guangdong Province in Southern China, pose a significant challenge for traditional wind resource assessment methods, which often rely on limited point measurements. These methods may not adequately capture the spatial variability of wind fields influenced by intricate terrain. This paper aims to address this gap by employing a sophisticated mesoscale modeling system to generate detailed, high-resolution wind resource maps for Guangdong, providing a more comprehensive understanding of its wind power potential.
Methods
The study utilizes a coupled mesoscale modeling system consisting of the PSU/NCAR Fifth-Generation Mesoscale Model (MM5) and the CALMET diagnostic meteorological model. MM5 is employed to generate a large-scale wind field as an initial guess, while CALMET refines these fields to a high resolution (down to 100m in Hong Kong) by incorporating detailed terrain and land-use data. This MM5/CALMET system was run on an hourly basis for three years (2004-2006) to simulate wind fields. The model's performance was validated against observed data from surface stations and upper-air soundings, showing good agreement. The Weibull probability distribution function was then used to characterize the wind speed variations and estimate wind power potential.
Datasets
The study used observational data from meteorological stations operated by the Hong Kong Observatory for model validation. Specifically, data from 41 meteorological surface stations and one upper-air sounding station (King's Park) were utilized. Additionally, data from three other meteorological surface stations not assimilated into the CALMET model were used for further validation. Field data from five stations, including EMSD, HLC, KSC, HHW, and XW, were also used for comparison. For initial and boundary conditions for the MM5 simulations, the 1.0-degree National Centers for Environmental Prediction (NCEP) Final Analyses (FNL) were used.
Key Insights and Ideas
- Coastal Dominance: Wind power potential is significantly higher in coastal areas of Guangdong compared to inland regions, driven by more consistent and stronger wind speeds.
- Weibull Distribution is Key: Simply using mean wind speed is insufficient for accurate wind power assessment. The Weibull distribution, which captures the variability of wind speeds (shape and scale parameters), is essential for a reliable estimation of wind power density.
- High-Resolution Modeling Matters: For complex terrains, high-resolution mesoscale modeling (like the MM5/CALMET system used here) is crucial to accurately represent wind field variations and avoid underestimating wind resources.
- Guangdong's Potential: The study identifies specific coastal locations like Zhanjiang, Shenquan, and Jinghai as having high wind resource potential suitable for wind farm development.
Key Results
- Approximately 70% of the time, wind speeds along the coast of Guangdong exceed 5 m/s, which is considered a baseline for typical wind turbines.
- The model simulations showed good agreement with observed wind data, with a mean RMSE of 0.93 and a mean Index of Agreement of 0.78 for wind speed.
- The shape parameter (k) of the Weibull distribution in coastal regions of Guangdong is generally above 2, indicating stable wind conditions favorable for wind turbines.
- The scale parameter (c) also shows significant variation in coastal areas, suggesting good wind power potential but also considerable year-to-year variability (up to 40-50%).
- Wind power per unit area is generally concentrated along coastal areas, with highest values (500-600 W/m²) observed along the coast.
Limitations and Future Work
The authors acknowledge that the CALMET model's performance is dependent on the density, frequency, and accuracy of observational data used as input. While the MM5/CALMET system with high resolution demonstrated good performance, the accuracy of the results is ultimately tied to the accuracy of the initial guess field from the MM5 simulation. Future work suggested includes incorporating an urban canopy model to better represent radiative exchange within complex urban environments, which could further improve the accuracy of meteorological fields, particularly in areas where wind turbines might be located near urban canyons. Continuous improvements to the system are recommended.
Summary updated