A Projection of Extreme Precipitation Based on a Selection of CMIP5 GCMs over North Korea

Jang Hyun Sung, Minsung Kwon, Jong‐June Jeon, Seung Beom SeoSustainability 2019 · DOI

TL;DR

This study tackles the challenge of selecting representative climate change scenarios for impact assessments by using the Katsavounidis-Kuo-Zhang (KKZ) algorithm to identify a subset of CMIP5 Global Climate Models (GCMs) that capture the full range of variability. The findings project a significant increase in the magnitude and frequency of extreme precipitation events in North Korea by the late 21st century, offering crucial insights for water resource management.

What Problem Does This Paper Address?

Assessing the impacts of climate change, particularly extreme events like heavy rainfall, is crucial for effective adaptation and mitigation strategies. However, decision-making is hampered by the sheer number of available climate change scenarios derived from numerous Global Climate Models (GCMs). Each GCM has its own strengths and weaknesses, and using all of them for impact assessments is computationally expensive and complex. Previous approaches often focused on model performance in reproducing historical climate, which doesn't guarantee future projection accuracy. This paper addresses the need for a more efficient yet robust method to select a representative subset of GCMs that accurately reflects the uncertainty and variability present across the entire ensemble, specifically for projecting extreme precipitation in North Korea.

Methods

The core methodological innovation in this study is the application of the Katsavounidis-Kuo-Zhang (KKZ) algorithm to select Representative Climate Change Scenarios (RCCS). Unlike methods that solely rely on historical performance, the KKZ algorithm aims to select GCMs that preserve the inter-model variability of the entire ensemble. The algorithm works by iteratively selecting models that best represent the spread of the data, ensuring that the chosen subset captures the range of potential future climate outcomes.

For the analysis of extreme precipitation, the study employed a Multi-Model Ensemble (MME) approach. They used daily precipitation data from 25 CMIP5 GCMs, downscaled to a local scale for 27 weather stations in North Korea using the SDQDM (Spatial Disaggregation with Quantile Delta Mapping) method. The Generalized Extreme Value (GEV) distribution was then used to model extreme precipitation events, estimating parameters like location, scale, and shape to characterize the magnitude and frequency of these events. The study projected changes for three future periods: F1 (2011-2040), F2 (2041-2070), and F3 (2071-2100), under the high-emission RCP 8.5 scenario.

Datasets

The study utilized:

  • Observed daily precipitation data: Collected from 27 meteorological stations in North Korea for the reference period (1980-2005).
  • CMIP5 Climate Projections: Daily precipitation data from 25 GCMs under the RCP 8.5 scenario. These projections were statistically downscaled to the local weather station level.

Key Insights and Ideas

  • The KKZ algorithm is a powerful tool for scenario selection: By focusing on preserving inter-model variability rather than just historical accuracy, the KKZ algorithm provides a more reliable way to select a subset of GCMs that can represent the full spectrum of potential future climate changes. This is crucial for managing computational resources while still accounting for uncertainty.
  • North Korea faces a significant increase in extreme precipitation: The projections indicate a clear trend of increasing magnitude and frequency of extreme precipitation events throughout the 21st century, particularly in the later periods. This highlights the vulnerability of the region to climate change impacts.
  • RCCS effectively represent the ensemble: The study demonstrated that the selected RCCS performed comparably to the full ensemble of all CCS in terms of simulating observed precipitation and projecting future changes. This validates the approach of using a smaller, representative set of scenarios.
  • Regional variations in projected changes: While extreme precipitation is projected to increase across North Korea, certain regions, particularly those near the MDL and along the Chuncheon River, are expected to experience more pronounced increases.

Key Results

  • RCCS capture ensemble variability: The study found that the selected RCCS successfully retained the full range of variability and the same mean value across all months as the full ensemble of all CCS.
  • Projected increase in extreme precipitation: For the period 2070-2099 (F3), extreme precipitation (defined as daily maximum precipitation with a 20-year return period) was projected to occur once every 8.3 years, a significant increase compared to the reference period.
  • Magnitude and frequency increase: The 20-year precipitation frequency was projected to increase by 5.2% in F1, 27.9% in F2, and 39.9% in F3 compared to the present.
  • GEV parameter shifts: Location and scale parameters of the GEV distribution increased from the present to F3, while the shape parameter decreased, indicating an increase in the magnitude and frequency of extreme events.
  • Return period acceleration: The spatially averaged return period for 20-year extreme precipitation was projected to decrease from 20.7 years in F1 to 11.7 years in F2 and 8.3 years in F3.

Limitations and Future Work

  • Focus on RCP 8.5: The study exclusively used the RCP 8.5 scenario, which represents a high-emission pathway. Future work could explore other RCP scenarios to provide a broader range of potential outcomes.
  • Statistical Downscaling: While SDQDM is a robust method, all statistical downscaling techniques have inherent limitations. Further research could explore dynamical downscaling for a more direct physical representation.
  • Flood Volume Projections: The paper suggests that future work could utilize the RCCS to evaluate the prospects of changes in flood volumes in North Korea, which would be a valuable extension for water resource management.
  • Adaptive Strategies: The study highlights the potential impacts of increased extreme precipitation, such as river erosion and sedimentation, and suggests the need for appropriate adaptive plans and cooperative adaptation strategies, particularly for shared watersheds.

Summary updated

Abstract

The numerous choices between climate change scenarios makes decision-making difficult for the assessment of climate change impacts. Previous studies have used climate models to compare performance in terms of simulating observed climates or preserving model variability among scenarios. In this study, the Katsavounidis-Kuo-Zhang algorithm was applied to select representative climate change scenarios (RCCS) that preserve the variability among all climate change scenarios (CCS). The performance of multi-model ensemble of RCCS was evaluated for reference and future climates. It was found that RCCS was well suited for observations and multi model ensemble of all CCS. Using the RCCS under RCP (Representative Concentration Pathway) 8.5, the future extreme precipitation was projected. As a result, the magnitude and frequency of extreme precipitation increased towards the farther future. Especially, extreme precipitation (daily maximum precipitation of 20-year return-period) during 2070-2099, was projected to occur once every 8.3-year. The RCCS employed in this study is able to successfully represent the performance of all CCS, therefore, this approach can give opportunities managing water resources efficiently for assessment of climate change impacts.
BibTeX
@inproceedings{sung2019projection,
  author = {Jang Hyun Sung and Minsung Kwon and Jong‐June Jeon and Seung Beom Seo},
  title = {{A Projection of Extreme Precipitation Based on a Selection of CMIP5 GCMs over North Korea}},
  booktitle = {Sustainability},
  year = {2019},
  doi = {10.3390/su11071976},
}
Figure