Changes in precipitation extremes in a warming climate

Changes in precipitation extremes in a warming climate Pao-Shin Chu Department of Meteorology School of Ocean & Earth Science & Technology University of Hawaii-Manoa Honolulu, HI, U. S. A.

IPCC AR 4 (2007) Giambelluca et al. , (2008)

Extreme events in temperature and precipitation such as summer heat waves, cold spells in winter, heavy rainfall/flooding, and drought are changing over time The occurrence of extreme events is a serious concern for society because of their potential damage to humans, property, public infrastructure, agriculture, transportation, and others. To better monitor and understand the variations of extreme events, the Climate Variability and Predictability (CLIVAR) program has developed a suite of climate change indices for a standard comparison.

• In this study, five of the relevant climate change indices suggested by WMO/WCRP/CLIVAR are applied to Hawaii daily precipitation data to investigate the possible changes of extremes. • Their relationships with the El Nino-Southern Oscillation (ENSO) and the Pacific Decadal Oscillation (PDO) are examined.

v. Observational data: COOP data from NCDC water year : July to June of the next year winter season: November through April of the next year v In order to maintain data quality, some criteria are applied to the data sets. 1. A month is considered as having complete data if there are 5 missing days. 2. A year is considered as complete if all months are complete according to (1). 3. A station series is considered as complete if it has 65% complete years according to (2). COOP stations Numbers The 1950 s-2007 The 1960 s-2007 The 1970 s-2007 The 1980 s-2007 37 41 50 65

Definition of the five indices Perspective Index Definition Unit Intensity SDII Average precipitation intensity in wet days mm/day Frequency R 25 Annual total number of days with precipitation 25. 4 mm days Magnitude R 5 d Annual maximum consecutive 5 -day precipitation amount mm Magnitude R 95 p Fraction of annual total precipitation due to events % exceeding the 1961 -90 95 th percentile Drought CDD Annual maximum number of consecutive dry days

Ø Mann-Kendall test and Sen’s method Ø A two-sample Kolmogorov-Smirnov (K-S) test Ø Fisher-Z transformation Ø Nonparametric Mann-Whitney test (Tu, Chou, and Chu, 2009, J. Climate, The abrupt shift of typhoon activity in the vicinity of Taiwan …)

Nonparametric Mann-Kendall test and Sen’s method • Mann-Kendall test assumes that the time series dataset obeys the model: • For data pair xj and xk, where j>k, the sign is calculated: • The statistics S is calculated: • If n 10, the normal approximation statistics Y, which is based on S will be calculated. • Positive S or Y means positive trend, negative S or Y means negative trend.

Nonparametric Mann-Kendall test and Sen’s method • When using Sen’s method to estimate the slope of the trend, first assume that f(t) in can be represented by: where Q is the slope to be estimated and B is a constant. • The slopes of all data pairs are calculated using where j>k. The median of all these slopes of data pairs is the Sen’s estimator of slope. • Mann-Kendall method tests whether the trend is increasing or decreasing and estimates the significance of the trend. • Sen’s method quantifies the slope of this trend. • Missing values are allowed in these two methods, and the data need not conform to any particular distribution. Besides, the Sen’s method is not greatly affected by single data errors or outliers.

HRI: Hawaii Rainfall Index, 27 gages from 3 islands. Data have been standardized. • Using monthly precipitation data, Chu and Chen (J. Climate, 2005) suggested there is a downward trend in Hawaiian rainfall during the last century.

In addition to the significance test applied to individual stations, field significance is tested. • In a given dataset, one would expect a certain number of stations or grids to pass a significance test at random. To ensure the significance at individual stations is not due to random chance, multiple testing is performed to investigate the field significance (multiplicity problem). • Assuming spatial independence, a binomial probability distribution is used to evaluate the overall significance of the trends.

Intensity Frequency 50 s: the 1950 s to 2007 60 s: the 1960 s to 2007 Etc. Magnitude Drought field significant at the 5% level

trends from the 1950 s to 2007 • Downward trends in SDII, R 25, and R 5 d for Kauai and Oahu • Upward trends in Hawaii Intensity Frequency

trends from the 1950 s to 2007 • For CDD, overall upward trends. Most islands tend to show longer, consecutive periods of no precipitation days since 1950 s. Magnitude Drought

Ø 30 -year running series are considered. Time-dependent changes in extreme rainfall events are examined for a 30 -yr interval, but the series are moving forward one year at a time. • 1950 -1979, 1951 -1980, until 1978 -2007 (29 series). Ø Analyze the trends of every 30 -year window. Ø Calculate the time derivatives of the trends (this allows us to determine whether the trends are stable or changing with time).

SDII Time series at Honolulu International Airport trends of 30 -year running series Trend of 1950 -1979 is plotted in the 1964, trend of 1951 -1980 is plotted in the 1965, etc. §Long-term downward trend §positive time derivative of trend ØLong term trend is changing in time (gentle or even changing sign).

CDD Time series at Opihihale trends of 30 -year running series §Long-term upward trend §positive time derivative of trend ØLong term trend is stable.

Intensity Frequency • Positive time derivatives and long-term downward trends in Kauai and Oahu. • Negative time derivatives and upward trends in Hawaii. Ø Trends are changing.

Magnitude Drought • For CDD, positive time derivatives and positive trends. Ø Trends are stable.

• Chu (1995) proposed a mechanism that associates deficient precipitation in Hawaii during winter with El Niño. • Chu and Chen (2005) suggested deficient precipitation in El Niño /+PDO phase, while abundant precipitation during La Niña /-PDO phase in Hawaii using monthly precipitation data.

The N-S vertical cross section of the meridional wind component and the negative pressure vertical velocity between 150°W-165° at 11 standard pressure levels for winter

• Positive correlations between four precipitation-related indices and SOI, and negative correlations between CDD and SOI (the Fisher Z transformation is applied to the original correlation coefficients). • For La Niña event (large and positive SOI), Hawaii not only tends to have more seasonal rainfall, but also receives more frequent heavy rainfall. For El Niño years, there are fewer extreme events. For CDD, shorter annual maximum consecutive dry days during La Niña events while they are longer for El Niño years.

Return periods of rain storms • Assessing the vulnerability of a region to extreme rainfall and associated flood events is an important step in disaster prevention plans. • The Generalized Extreme Value (GEV) distribution and its L-moment estimation

ξ : location, α : scale, ĸ : shape

Oahu

• Previous analysis is based on a stationary GEV model by assuming climate does not change (stationary). Recently we have applied a nonstationary GEV method so that the parameters are allowed to vary with time. Return levels are changing with time.

Dynamical downscaling for future rainfall variations for an island (C. Norton) • IPCC AR 4 GCMs • WRF 3. 2 model Two-way nesting, very high-resolution (1 km) resolution (super-computer) Noah land surface CAM radiation scheme (NCAR community Atmosphere model) Yonsei University boundary layer scheme Tiedtke cumulus parameterization Cloud microphysics (WSM 6)

Domain 4 nested: Parent: 50 km 2 nd: 25 km 3 rd: 5 km 4 th: 1 km 28 Vertical layers

January rainfall for 2030 and 2040

Summary • Trends of five climate change indices related to extreme precipitation events in Hawaii are investigated using daily observational records from the 1950 s to 2007. • A nonparametric rend analysis suggests long-term downward trends for four precipitation-related indices, and long-term upward trends for CDD. • Time-dependent changes in extreme precipitation events are examined for a 30 -yr interval. • A non-stationary GEV model is applied to examine trends in return levels.

• The time derivatives of trends of the 30 -yr running series imply that there is a phase change for all 4 precipitation-related series. Since 1980 s, there seems to be an increase in precipitation intensity, frequency, and magnitude of intense events in Oahu and Kauai. For CDD, the long-term increasing trend is stable, implying a stable lengthening of annual maximum consecutive dry days. • Positive relationships are found between the wetness indices and SOI, and negative relations between CDD and SOI. This suggests more precipitation extremes during La Niña years and vice versa for El Niño years. • Return levels are changing with time. • Dynamical downscaling for future rainfall variations for an island. Be applied to Taiwan?

• Chen, Y. , P. -S. Chu, and T. Schroeder, 2011: Trends in precipitation extremes and return levels in the Hawaiian Islands under a changing climate. In preparation. • Chu, P. -S. , Y. Chen, and T. Schroeder, 2010: Changes in precipitation extremes in the Hawaiian Islands in a warming climate. J. Climate, 23, 4881 -4900. • Chu, P. -S. , X. Zhao, Y. Ruan, and M. Grubbs, 2009: Extreme rainfall events in the Hawaiian Islands. J. Appl. Meteor. Climatol. , 48, 502 -516. • Chu, P. -S. and H. Chen, 2005: Interannual and interdecadal rainfall variations in the Hawaiian Islands. J. Climate, 18, 4796 -4813.

Thank You!