CCI-HYDR Perturbation Tool. A climate change tool for generating perturbed time series for the Belgian climate MANUAL, JANUARY 2009

Transcription

1 CCI-HYDR project (contract SD/CP/03A) for: Programme SSD «Science for a Sustainable Development» MANUAL, JANUARY 2009 CCI-HYDR Perturbation Tool A climate change tool for generating perturbed time series for the Belgian climate Faculty of Engineering Department of Civil Engineering Hydraulics Division CCI-HYDR project Royal Meteorological Institute of Belgium Meteorological Research and Development Department Risk Analysis and Sustainable Development Section

2 Faculty of Engineering Department of Civil Engineering Hydraulics Section Kasteelpark Arenberg 40 BE-3001 Leuven, Belgium tel fax Meteorological Research and Development Department Risk Analysis and Sustainable Development Section Avenue Circulaire, 3 BE-1180 Brussels, Belgium tel fax Emmanuel.Roulin@oma.be No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without indicating the reference : Ntegeka V., Willems P., CCI-HYDR Perturbation Tool: a climate change tool for generating perturbed time series for the Belgian climate, Manual version January 2009, K.U.Leuven Hydraulics Section & Royal Meteorological Institute of Belgium, 7 p.

3 CCI-HYDR Perturbation Tool i Table of contents 1 Introduction How are the scenarios defined? What does the program do? What temporal and spatial scales are we looking at? What time horizons are provided? What does the output series represent? What variables are considered in the program? Procedure for running the program Limitations of the scenarios References... 7

4

5 CCI-HYDR Perturbation Tool 1 1 Introduction There is a dearth of climate change transformation tools especially for Belgium. This is in part explained by the lack of climate change scenarios relevant for local regions. With the completion of the PRUDENCE project in 2004, new regional climate scenarios became available. However, the number of climate models coupled with emission scenarios made it less obvious for users to synthesize. The CCI-HYDR project on Climate change impact on hydrological extremes in Belgium for the Belgian Science Policy Office (Programme Science for a sustainable development ) was setup to primarily synthesize the pertinent climate scenarios for Belgium. The project opted to focus on studying the high resolution regional climate models from the PRUDENCE project. One of the objectives required providing end users with time series which would enable them to incorporate the impacts of climate change within their future water management plans. The water authorities in particular were singled out as one of the target end users of the scenarios. A perturbation algorithm was developed so that impact analysts in Belgium can assess the hydrological impacts of climate change. The algorithm imparts a perturbation to an observed series to generate future time series. The observed time series are perturbed on the basis of four SRES scenarios (A1B, A2, B1 and B2). The climate model simulations with the A2 and B2 regional scenarios were extracted from the PRUDENCE database while the A1B and B1 scenarios were extracted from the IPCC AR4 database. Most hydrological climate change impact assessments require inputs of rainfall and potential evapotranspiration (ETo) to generate runoffs. By comparing the historical runoffs to the future runoffs impacts can be quantified. Moreover, the scenarios have combined several model outputs into only three scenarios that reliably represent the overall range of expected impacts. The earlier version of the CCI-HYDR Perturbation Tool only involved scenarios for rainfall and ETo, while the current version includes scenarios for temperature and wind speed, developed by the project Klimaatscenario s voor Vlaanderen for INBO (Instituut voor Natuur- en Bosonderzoek) (Demarée et al., 2008; Baguis et al., 2008). 2 How are the scenarios defined? For ease of interpretation, three scenarios are defined as high, mean and low based on the expected hydrological impacts. The definition is not dependent on the projected rainfalls alone. Rather it is based on the combined effect of the rainfall and ETo; in other words, the variables are combined to generate an impact which can then be classified as high, mean and low. The average temperature and wind speed variables are inferred through a correlation analysis with rainfall and ETo. Wind speed is positively correlated with rainfall during winter and negatively correlated during summer. The average temperature is assumed to follow the ETo trends, i.e., ETo is directly dependent on temperature. The high scenario projects a future with wet winters and dry summers while the low scenario projects a future with dry winters and dry summers. Thus, it is expected that the risk associated with flooding is higher in the high scenario than the low scenario which is critical for low flows. In essence, the high, mean and low scenarios may be referred to as wet, mild and dry scenarios respectively. It is notable that the mean scenario represents mean conditions and is not the best future guess. Users are encouraged to consider all the three scenarios to account for the overall uncertainty. Figure 1 illustrates the relevance and interpretation of the scenarios.

6 2 CCI-HYDR Perturbation Tool High/Wet Mean/Mild Low/Dry Hydrological Impact Floods Low flows Figure 1: Relevance and interpretation of the CCI-HYDR scenarios. 3 What does the program do? The program perturbs or changes the input series of rainfall (mm), ETo (mm), temperature (ºC) and wind speed (m/s). For rainfall, the changes in number of wet days and intensities are considered. The changes are quantile based to account for the fact that the changes might depend on the magnitude or return period of the event. ETo is transformed in a relatively easier way by applying seasonal average changes because high quantiles tend to have similar percentage changes to lower quantiles. Wind speed is generated by using the correlation with precipitation. High wind speeds in winter tend to be correlated with high rainfall amounts; however during summer, high speeds imply dry conditions. Temperature is treated in a similar way with ETo meaning high ETo values imply high temperatures; low ETo rates are outcomes of low temperatures. More details about the perturbation procedure can be found in chapter five of the CCI-HYDR Technical Report II Study of climate change scenarios (see project website: 4 What temporal and spatial scales are we looking at? The software uses time series at hourly and daily time steps. These are time scales relevant for river subbasins. The scenarios were developed mainly for catchments up to 1000 km 2. 5 What time horizons are provided? The time series perturbation procedure was developed from the PRUDENCE regional climate models which mainly dealt with a 30 year control period of and a 30 year scenario period of Interpolation is made for other periods to account for potential differences between the period covered by the input series and the standard control period. The use of a 30 year period would be ideal given that it covers exactly the length of the climate model control and scenarios periods. Also, a 30 year period corresponds to an average climate oscillation cycle (Ntegeka and Willems, 2008). Perturbed series can also be conducted for shorter or longer periods, but the rainfall-runoff modeller has to keep in mind that the results may be biased from the long-term averaged climate. Due to the oscillations, the impact will be overestimated if the input series period covers an oscillation peak and underestimated when it covers an oscillation low.

7 CCI-HYDR Perturbation Tool 3 6 What does the output series represent? The output series is the perturbed input series for a given time horizon in the future. Target years of 2020, 2030, 2040, 2050, 2060, 2070, 2080, 2090, and 2100 can be selected. Each target year is in the centre of a 30 year block e.g., 2050 represents changes from Thus, no predictions are made for a particular year or day in the future. It is important to emphasise that the future changes are most reliable if the observed data input is 30 years and covering the period Since the modelled future scenario period was , the changes during the target years within the block, i.e., 2070, 2080 and 2090, would also be more reliable. For other target years, the interpolation and extrapolation of the changes leads to less certain future perturbations. 7 What variables are considered in the program? The software was originally designed with the key aim of aiding hydrological impact assessments. However, temperature and wind speed have also been included for end users to investigate other possible impacts. Rainfall and ETo are the main parameters required by conceptual hydrological models to investigate impacts. Although it is possible to run the program using only one of the variables, it is preferred that al least rainfall and ETo variables are run in a hydrological model to study the impacts. Studying the changes of one variable requires an understanding of the methodology of the scenario development provided in chapter five of the CCI-HYDR Technical Report II Study of climate change scenarios (see project website: 8 Procedure for running the program The program uses Excel VBA macros which need permission to run in Excel. The user permits the program to run by first changing the macro security options. This is done by using Tools Macro Security Security Level Medium. On opening the program, the user clicks enable macro and a window appears requesting the user to open the file as a read only. Accept and close the logo shown as in Figure 2 (click the X ). Figure 2: Startup window for the program.

8 4 CCI-HYDR Perturbation Tool For daily data, the user pastes the rainfall and ETo data beginning in row 4 in the section of the input series (Figure 3). The formats for the dates are indicated in the sheet. It is possible to only insert either rainfall or ETo alone but for hydrological modelling the two variables are necessary. Make sure that the daily data option is selected. The region for the station is selected i.e., coastal or inland. The coastal area in Belgium was found to show some differences in the projected changes from the inland regions (Ntegeka et al., 2008). The baseline is the period that is considered to be a reference for the future climate change. This baseline is taken as the period defined by the climate model control period. For the climate models, the is taken as the baseline. Therefore if a period s average climate is not significantly different from the period, then the option for Baseline ( ) is selected from the dropdown list. It is also possible for the baseline to be taken as defined from the input data period; here there is need interpolation to account for the difference in time i.e., It is recommended that the option Baseline (from input) be selected always. After selecting a baseline option, one can then select a target year from the drop-down list and then click Perturb. The program then begins to generate the perturbed series. On completion, the daily perturbed series will be shown in the perturbed series section with the target year shown in the title. Figure 3: Input of daily rainfall and ETo series (dummy data). The perturbed series (Figure 4) represent the changed input series, where the perturbation corresponds to the climate change from the current climate (based on input period) to the predicted future climate (based on the target year). For each variable, 3 perturbed series are produced with each representing a scenario i.e., high, mean and low. If a variable was not provided, the time series will be blank. The perturbed series can then be copied and pasted into another excel workbook or in a text file and stored.

9 CCI-HYDR Perturbation Tool 5 Figure 4: Generated daily data scenarios (dummy data). If the hourly or 10min data option is selected, the user is required to prepare the input text file before clicking Perturb. Two text files are required with one containing the date information and the other containing the values. The hourly or 10 min option is only available for rainfall data. ETo data if available should be pasted in the sheet before clicking the Perturb option. It is not necessary to paste the rainfall daily data in the sheet if the hourly option is selected. One can clear the data using the Clear data button, insert the daily ETo, select the hourly or 10 min option and then click Perturb. The window shown in Figure 5 appears. Figure 5: Hourly data input. If you have not yet prepared the hourly input data click cancel to exit. If you had prepared the input text files then browse and select the text files. Make sure that the rainfall text file is not wrongly inserted as the time file. The formats for the hourly time and rainfall are shown in Table 1. The rainfall text file contains the rainfall values in a single column without headers and similarly the hourly time text file contains the hourly time data in a single column. Using text files often leads to errors if the text files for rainfall and time have different number of lines. It is advisable to remove empty lines in the text files and to check that both rainfall and time files have the same number of lines (put cursor at end of text file and use Edit Goto in notepad).

10 6 CCI-HYDR Perturbation Tool Table 1: Text formats for hourly/10min data input Time text file format Rainfall text file format 01/01/ :00: /01/ :00: /01/ :00: /01/ :00: /01/ :00: /01/ :00: Clicking Run will begin the perturbation process for the hourly/10min data. This, however, takes longer than the daily input. On completion, three time series are found in the stated output folder (Figure 5). The files are named hourly_high.txt/10min_high.txt, hourly_mean.txt/ 10min_mean.txt and hourly_low.txt/10min_low.txt. These files can then be used for further analysis. Note that if the present ETo data was pasted in the sheet, the future ETo time series information will be provided in the sheet and not in a text file. These files can then be renamed as they are replaced with future runs. To rerun the program the user first clears the previous data using the Clear data button. In case of any errors, bugs or suggestions, the user may contact the program developers. The contact information can be found in the Perturbation option in the menu bar. One required windows setting is that decimal points are taken as decimal points and not commas. This can be checked in windows by changing the control panel language settings to English UK or another appropriate setting. For missing values, the user is expected to infill the time series before running the program. 9 Limitations of the scenarios The changes embedded in the program were based on 30 year time scales. This implies that longer time scales required for high return periods are not directly available. However they may be extrapolated from the 30 year time periods albeit with extrapolation assumptions. The scenarios are more reliable for the target years within the period and for input series within the control period Instead of interpolating and extrapolating, other periods require more transient runs to increase confidence in the projections. The climate scenarios were primarily based on the PRUDENCE regional climate models which were mainly driven by one Global Circulation Model (HadAM3H A2). This precluded a robust estimation of the uncertainties. Successive projects like the ENSEMBLES project are including more emission scenarios and time horizons at regional scales; this may lead to future updates.

11 CCI-HYDR Perturbation Tool 7 10 References CCI-HYDR project reports and papers: Ntegeka V., Willems P., Baguis P., Roulin E., Climate change impact on hydrological extremes along rivers and urban drainage systems. Summary report Phase 1: Literature review and development of climate change scenarios, K.U.Leuven Hydraulics Section & Royal Meteorological Institute of Belgium, April 2008, 64 p. Baguis P., Boukhris O., Ntegeka V., Roulin E., Willems P., Demarée G., Climate change impact on hydrological extremes along rivers and urban drainage systems. I. Literature review, Technical report, K.U.Leuven Hydraulics Section & Royal Meteorological Institute of Belgium, May 2008, 57 p. Ntegeka V., Baguis P., Boukhris O., Willems P., Roulin E., Climate change impact on hydrological extremes along rivers and urban drainage systems. II. Study of rainfall and ETo climate change scenarios, Technical report, K.U.Leuven Hydraulics Section & Royal Meteorological Institute of Belgium, May 2008, 112 p. Ntegeka V., Willems P., Climate change impact on hydrological extremes along rivers and urban drainage systems. III. Statistical analysis of historical rainfall, ETo and river flow series trends and cycles, Technical report, K.U.Leuven Hydraulics Section & Royal Meteorological Institute of Belgium, May 2008, 37 p. Ntegeka, V., Willems P., Trends and multidecadal oscillations in rainfall extremes, based on a more than 100-year time series of 10 min rainfall intensities at Uccle, Belgium, Water Resour. Res., 44, W07402, doi: /2007wr INBO project reports: Demarée G., Baguis P., Deckmyn A., Debontridder L., Pinnock S., Roulin E., Willems P., Ntegeka V., Kattenberg A., Bakker A., Lenderink G., Bessembinder J., Klimaatscenario s voor Vlaanderen, rapport voor het Instituut voor Natuur- en Bosonderzoek (INBO), Koninklijk Meteorologisch Instituut van België, Koninklijk Nederlands Meteorologisch Instituut (KNMI) & K.U.Leuven Afdeling Hydraulica, december Baguis P., Ntegeka V., Willems P., Roulin E., Extension of CCI-HYDR climate change scenarios for INBO, Technical report, K.U.Leuven Hydraulics Section & Royal Meteorological Institute of Belgium, January 2009, 31 p.

12

13 K.U.Leuven Faculty of Engineering Department of Civil Engineering Hydraulics Division Kasteelpark Arenberg 40 BE-3001 Leuven, Belgium tel fax Royal Meteorological Institute of Belgium Meteorological Research and Development Department Risk Analysis and Sustainable Development Section Avenue Circulaire, 3 BE-1180 Brussels, Belgium tel fax Emmanuel.Roulin@oma.be