1. Climate Extremes PRECIS Workshop Tanzania Meteorological Agency, 29 th June – 3 rd July 2015

2. Contents What is ‘extreme’? How can we calculate extremes? Observed changes in climate extremes Projections of extremes from IPCC Caution: validating extremes Why do extremes matter? An ‘extreme’ example: frequency of daily rainfall events in Bangladesh

3. Contents What is ‘extreme’? How can we calculate extremes? Observed changes in climate extremes Projections of extremes from IPCC Caution: validating extremes Why do extremes matter?

4. The term ‘extreme’ can be used to describe phenomena on a wide range of space and time scales –i.e. heavy rainfall events (small scale) to widespread drought (large scale) An ‘extreme’ event can be associated with: –High impact events –Unprecedented events (in the available record) –Rare events (long return periods) –Exceedance of a threshold or percentile level in daily temperature or precipitation –Persistence of weather conditions –Climatic extremes (e.g. extreme seasons) What is ‘extreme’?

6. Internationally coordinated core set of 28 descriptive indices describe frequency, amplitude, and persistence of moderate extremes http://etccdi.pacificclimate.org/ Expert Team on Climate Change Detection and Indices (ETCCDI) How can we calculate ‘extremes’?

7. Indices – temperature based APPENDIX A: List of ETCCDMI core Climate Indices ID Indicator name DefinitionsUNITS FD0Frost daysAnnual count when TN(daily minimum)<0 º CDays SU25Summer daysAnnual count when TX(daily maximum)>25 º CDays ID0Ice daysAnnual count when TX(daily maximum)<0 º CDays TR20Tropical nightsAnnual count when TN(daily minimum)>20 º CDays GSL Growing season Length Annual (1st Jan to 31 st Dec in NH, 1 st July to 30 th June in SH) count between first span of at least 6 days with TG>5 º C and first span after July 1 (January 1 in SH) of 6 days with TG<5 º C Days TXxMax TmaxMonthly maximum value of daily maximum temp ºCºC TNxMax TminMonthly maximum value of daily minimum temp ºCºC TXnMin TmaxMonthly minimum value of daily maximum temp ºCºC TNnMin TminMonthly minimum value of daily minimum temp ºCºC TN10pCool nightsPercentage of days when TN<10th percentileDays TX10pCool daysPercentage of days when TX<10th percentileDays TN90pWarm nightsPercentage of days when TN>90th percentileDays TX90pWarm daysPercentage of days when TX>90th percentileDays WSDI Warm spell duration indicator Annual count of days with at least 6 consecutive days when TX>90th percentile Days CSDI Cold spell duration indicator Annual count of days with at least 6 consecutive days when TN<10th percentile Days DTR Diurnal temperature range Monthly mean difference between TX and TN ºCºC

8. Indices – precipitation based APPENDIX A: List of ETCCDMI core Climate Indices IDIndicator nameDefinitionsUNITS RX1day Max 1-day precipitation amount Monthly maximum 1-day precipitationMm Rx5day Max 5-day precipitation amount Monthly maximum consecutive 5-day precipitationMm SDII Simple daily intensity index Annual total precipitation divided by the number of wet days (defined as PRCP>=1.0mm) in the year Mm/day R10 Number of heavy precipitation days Annual count of days when PRCP>=10mmDays R20 Number of very heavy precipitation days Annual count of days when PRCP>=20mmDays Rnn Number of days above nn mm Annual count of days when PRCP>=nn mm, nn is user defined threshold Days CDD Consecutive dry days Maximum number of consecutive days with RR<1mmDays CWD Consecutive wet days Maximum number of consecutive days with RR>=1mmDays R95p Very wet days Annual total PRCP when RR>95 th percentileMm R99p Extremely wet days Annual total PRCP when RR>99 th percentilemm PRCPTOT Annual total wet-day precipitation Annual total PRCP in wet days (RR>=1mm)mm

9. Example: Calculating TX90p (warm days) Calculate threshold exceeded by the 10% hottest days (Tmax) in the present-day or ‘baseline’ period (i.e. 1961-90) On average, in the baseline period, 10% of days (approx. 36) will exceed this threshold 10% days exceed 23.2º (~ 36 days) 23.2º 19611990

10. Calculate the average number of times that same threshold is exceeded in a future period In this future period, 58% of days (212 days) now exceed this threshold! 23.2º 20702100 58% days exceed 23.2º (av. 212 day) (NOTE: these are synthetic data, not from real projections!) Example: Calculating TX90p (warm days)

11. R95PTOT- Total annual rainfall on heavy rain days Similarly for rainfall, we can define the number of ‘wet days’ (i.e. Days >1mm), and calculate the 95 th percentile of wet days (i.e. the 5% wettest ‘wet days’) in the baseline period These are‘ heavy rainfall days’ We can use this information to calculate the average amount of rainfall per year that occurs in ‘heavy’ events. 12.7 mm 19611990

12. Using the same threshold as before, we can now identify the ‘heavy’ rainfall days in the future This again allows us to calculate the average amount of ‘heavy’ rainfall per year in the future period, which in the example below will be a higher amount than what we found for the present day 20702100 12.7 mm R95PTOT- Total annual rainfall on heavy rain days

13. Observed Changes in Climate Extremes IPCC AR4 WG1, adapted from Alexander et al. (2006)

14. Projections of Climate Extremes (from IPCC AR5)

15. Added value from using RCMs Using GCMs alone to assess changes in extremes will often miss the influence of regional dynamics By running RCMs, we’re able to better detect the influence of complex topography and regional characteristics on extremes Fig 5. from Frei et al., 2006

16. From global to local climate …. Caution: Validating Extremes

17. Best compromise is to use gridded datasets (e.g. CRU/APHRODITE) to validate model output However also be aware of limitations of gridded data sets (e.g. density of station network) Reanalysis products are another useful option for validation, but can contain well-known biases which are generally a product of the reanalysis model Temperature reanalyses are typically more reliable than precipitation fields APHRODITE station network Caution: Validating Extremes

18. Why do ‘extremes’ matter?

19. An extreme example – Daily rainfall extremes in Bangladesh Bangladesh Phase 2 Project: high-resolution regional climate modelling for administrative regions in Bangladesh funded by UKAid within DfID collaborated with Bangladesh University for Engineering and Technology (BUET) used an ensemble modelling approach (PRECIS) to provide a wide range of climate change projections for the Bangladesh region

20. A percentiles approach to daily rainfall Used a percentiles-based approach to define various categories of daily rainfall events over Bangladesh 40-year observational dataset (APHRODITE v.10) Specifically, we defined the following categories: Low: 0 - 25 th percentile (0 - 6.1 mm/day) Moderate: 25 th - 75 th percentile (6.1 – 11.6 mm/day) Heavy: 75 th – 99 th percentile (11.6 – 21.4 mm/day) Very heavy: >99 th percentile (> 21.4 mm/day) Can now apply these thresholds to a present-day and a future 30-year time slice from our climate model data

21. Changes in daily rainfall events: 2070-2099 minus 1961- 2000

22. Impacts of projected changes in daily rainfall Shift towards more frequent heavy rainfall events suggests an increased risk in flash flooding events for the Bangladesh regions A decrease in the frequency of low-moderate rainfall events, combined with a decrease in the number of wet days during the monsoon season (not shown), could lead to more frequent dry spells, and increased risk of severe drought conditions Changes in extreme rainfall characteristics could have detrimental impacts on water resource management, agricultural productivity, and the overall well-being of the Bangladesh community.

23. Thank you! Q&A

24. © Crown copyright Met Office Model Validation