3 -rd Year Lab course Mineral desert dust

3 -rd Year Lab course Mineral desert dust: Comparison between Tel-Aviv University dust forecasts and PM 10 data Pavel Kishcha Tel-Aviv University email: pavelk@post. tau. ac. il

The Objectives of the Lab The main goal is to evaluate TAU 24 -hour dust forecasts by using Israel PM 10 data. This will help to better understand the model’s capabilities for providing reliable dust forecasts. PM 10 measurements taken at the Tel-Aviv University site (Yad Avner Building, 32. 1 N, 34. 8 E) will be used for model evaluation; To distinguish between different routes of Saharan dust transport into Israel by using the on-line archive of dust forecasts: from the eastern part of the Sahara, through Egypt, into Israel; or from the western part of the Sahara, through Southern Europe, into the Eastern Mediterranean and Israel; To estimate seasonal variations and the long-term tendency of desert dust events over Tel-Aviv, by using PM 10 data. Pavel Kishcha, Desert dust lab work

Lab Course Realization 1. Download daily PM 10 data from March 30, 1999, to the present taken at the Tel-Aviv site (Yad Avner Building). http: //www. sviva. gov. il/Doc. Generator. asp 2. Obtain seasonal variations of desert dust by analyzing monthly numbers of days with PM 10 ≥ 100 μg/m 3. 3. Obtain year-to-year variations of annual numbers of days with PM 10 ≥ 100 μg/m 3 and their linear fit, which illustrates the long-term tendency of desert dust over Tel-Aviv. 4. Select some examples of different routes of desert dust intrusions into Israel, by using the on-line archive of TAU dust forecasts: http: //wind. tau. ac. il/dust 8/YYMMDD/dust. html 5. Compare model-predicted surface dust concentrations with PM 10 data taken at the Tel-Aviv University site for two dust events in Israel, February 23 – 26, 2007, and March 10 – 14, 2007. 6. Compare model-predicted surface dust concentrations for all consecutive days from February to April 2006 with corresponding PM 10 data and estimate their correlation. Pavel Kishcha, Desert dust lab work

Geographical distribution of arid regions Sahara The Sahara is the world’s most important dust source; it is responsible for up to half of the global mineral dust emissions.

The dust storm over North Africa on February 23, 2006. The Moderate Resolution Imaging Spectroradiometer (MODIS) flying onboard the NASA Aqua satellite took this image. Pavel Kishcha, Desert dust lab work

The dust intrusion into the Atlantic Ocean on September 29, 2008. The MODIS image from the NASA Aqua satellite is used. Pavel Kishcha, Desert dust lab work

The Voyage of H. M. S. Beagle around the world (1831 – 1836) The Beagle “…The dust falls in such quantities as to dirty everything on board, and to hurt people’s eyes; vessels even have run on shore owing to the obscurity of the atmosphere. It has often fallen on ships when several hundred, and even more than a thousand miles from the coast of Africa. In some dust which was collected on a vessel, I was much surprised to find particles of stone above the Charles Darwin thousandth of an inch square, mixed with finer matter…” 1809 - 1882 [Charles Darwin, 1831]. Pavel Kishcha, Desert dust lab work

The Mediterranean Israeli Dust Experiment (MEIDEX) - winter 2003 The aim was to investigate the geographical variation of the optical, physical, chemical and mineralogical properties of desert aerosol particles transported over the Mediterranean. Numerical simulations of the transport of dust plumes were conducted with the purpose of determining the evolution of the dust transport and vertical distribution of Saharan dust layers. Ref. : Alpert, P. , S. O. Krichak, M. Tsidulko, H. Shafir, and J. H. Joseph, 2002, A dust prediction system with TOMS initialization. Mon. Weather Rev. 130, No. 9, 23 -35 -2345. Alpert, P. , Kishcha, P. , Shtivelman, A. , Krichak, S. O. , and Joseph, J. F. , 2004. Vertical distribution of Saharan dust based on 2. 5 -year model predictions. Atmospheric Research, 70, 109 -130. Pavel Kishcha, Desert dust lab work

In the year 2006, TAU Weather Research Center started running the modified version of the DREAM model (Nickovic et al. , JGR, 2001). TAU Dust Forecast Web page: http: //wind. tau. ac. il/dust 8/dust. html DREAM simulates all major processes of the atmospheric dust cycle; A detailed set of 8 particle size classes with effective size between 0. 1 and 7. 1 microns (0. 15, 0. 25, 0. 45, 0. 78, 1. 3, 2. 2, 3. 8, and 7. 1 m) is used; Dry (Georgi, 1986) and wet deposition processes are included. DREAM is initialized with the NCEP analysis and the lateral boundary data are updated every 6 h by the NCEP GFS model. The runs start at 1200 UTC and forecasts are performed for 3 -hour periods up to 72 hours ahead. Ref. : Kishcha et al. (JGR, 2005), Vertical distribution of Saharan dust over Rome (Italy): Comparison between 3 year model predictions and lidar soundings. Kishcha et al. (JGR, 2007), Forecast errors in dust vertical distributions over Rome (Italy): Multiple particle size representation and cloud contributions. Kishcha et al. , (2008), Saharan Dust over the Eastern Mediterranean: Model Sensitivity. Air Pollution Modeling and Its Application XIX, Springer. Pavel Kishcha, Desert dust lab work

Mean atmospheric lifetime for the eight size classes of mineral dust aerosols between 0. 1 and 10 m based on model results by Tegen and Lacis (JGR, 1996) Effective size, m 0. 15 0. 25 0. 40 0. 80 1. 50 2. 50 5 8 Lifetime, hours 231 229 225 219 179 126 67 28 The new DREAM model was capable of taking into account most of long-lived dust aerosols, in contrast to the old DREAM model. Pavel Kishcha, Desert dust lab work

Typical dust transport into Israel in spring (March 13, 2006): from the Sahara, through Egypt, into Israel Sea. WIFS satellite data TAU fcst Comparisons between PM 10 data in Tel. Aviv and 24 -h model-predicted data show that the TAU dust system produces dust forecast of acceptable accuracy. Pavel Kishcha, Desert dust lab work

Long-distance dust transport from the Western Sahara, through Southern Europe, into the Eastern Mediterranean between 4 - 9 April , 2006 Comparisons between PM 10 data in Tel-Aviv and 24 -h model-predicted data Pavel Kishcha, Desert dust lab work

Dust transport from Saudi Arabia into Israel on March 17, 2008. MODIS satellite data The DREAM-8 dust forecast Tel-Aviv University Weather Research Center (TAU We. RC) http: //wind. tau. ac. il/dust 8/dust. html Pavel Kishcha, Desert dust lab work

Comparisons between 24 -h DREAM-predicted dust concentration and PM 10 data at Tel-Aviv in March 2006 Correlation: R (DREAM-8) = 0. 76; R (DREAM orig) = 0. 65 Pavel Kishcha, Desert dust lab work

Interface of the on-line data base of the Israeli Ministry of Environmental Protection http: //www. sviva. gov. il Pavel Kishcha, Desert dust lab work

Pavel Kishcha, Desert dust lab work

Pavel Kishcha, Desert dust lab work

The long-term tendency of dust events over Israel (1999 – 2007) as inferred from year-to-year variations of annual numbers of dusty days with PM 10 greater than 100 g/m 3 PM 10 measurements at the Tel-Aviv University campus (Yad Avner Building) were used

Averaged seasonal variations of dust events over Israel (1999 – 2007) as inferred from monthly numbers of dusty days with PM 10 greater than 100 g/m 3 PM 10 measurements at the Tel-Aviv University’s campus (Yad Avner Building) were used

Tel-Aviv University desert dust forecasts are available via the Internet. http: //wind. tau. ac. il/dust 8/YYMMDD/dust. html where YY stands for year, MM for month, and DD for day. For example, to get the model-predicted dust transport on March 13, 2006, one can use the following link to the forecast of March 12, 2006 (the previous date): http: //wind. tau. ac. il/dust 8/060312/dust. html Pavel Kishcha, Desert dust lab work

Surface dust concentrations in the ascii format in Tel. Aviv are available by using the following link: http: //wind. tau. ac. il/dust 8/YYMMDD/surface-PM 10 -ramat-aviv-ascii. dat Here is some explanation of those files: Within each file you will see 13 numbers of predicted surface dust concentrations (X 1, X 2, …, X 13) started from 12: 00 UTC of the forecast day for consecutive intervals of 6 hours up to 72 hours ahead. # UTC Forecast Model time data [microgram/m^3] 1 12: 00 0 X 1 2 18: 00 6 X 2 3 24: 00 12 X 3 4 06: 00 18 X 4 5 12: 00 24 X 5 6 18: 00 30 X 6 7 24: 00 36 X 7 8 06: 00 42 X 8 9 12: 00 48 X 9 10 18: 00 54 X 10 11 24: 00 60 X 11 12 06: 00 66 X 12 13 12: 00 72 X 13 Pavel Kishcha, Desert dust lab work

How to compare modeled data with PM 10: an example for March 13, 2006. The model-predicted surface dust concentrations based on the model forecast of March 12, 2006 http: //wind. tau. ac. il/dust 8/060312/surface-PM 10 ramat-aviv-ascii. dat 2. 721 e+02 2. 370 e+02 2. 417 e+02 1. 740 e+02 2. 628 e+02 3. 248 e+02 3. 958 e+02 4. 371 e+02 6. 355 e+02 4. 843 e+02 2. 938 e+02 1. 396 e+02 7. 612 e+01 PM 10 data taken at the Tel-Aviv University site on March 13, 2006: יד אבנר - רמת אביב תאריךשעה (PM 10) חלקיקים נשימים מיקרוגרם/מטר קוב 13/03/2006 08: 00 29. 95 13/03/2006 14: 00 475. 07 13/03/2006 20: 00 369. 82 14/03/2006 02: 00 308. 73

Comparisons between PM 10 data in Tel-Aviv and 24 -h modelpredicted data from March 11 - 15, 2006. Pavel Kishcha, Desert dust lab work

The lab course realization step-by-step: 1. Download daily PM 10 data taken at the Tel-Aviv site (Yad Avner Building) from March 30, 1999, to the present. http: //www. sviva. gov. il/Doc. Generator. asp 2. Obtain seasonal variations of desert dust by analyzing monthly numbers of days with PM 10 ≥ 100 μg/m 3, averaged from 1999 – 2011. 3. Obtain year-to-year variations of annual numbers of days with PM 10 ≥ 100 μg/m 3 and their linear fit, which illustrates a long-term tendency of desert dust over Tel-Aviv during the period from 1999 - 2011. 4. Determine some examples of different routes of desert dust intrusions into Israel, by using the on-line archive of TAU dust forecasts: http: //wind. tau. ac. il/dust 8/YYMMDD/dust. html 5. Compare model-predicted surface dust concentrations with PM 10 data taken at the Tel-Aviv University site for two dust events in Israel, February 23 – 26, 2007, and March 10 – 14, 2007. 6. Compare model-predicted surface dust concentrations for all consecutive days from February to April 2006 with corresponding PM 10 data and estimate a correlation coefficient. Pavel Kishcha, Desert dust lab work

Report A report should include five Excel work sheets with the following five figures and their data: Fig. 1. Seasonal variations of desert dust over Tel-Aviv based on monthly numbers of days with PM 10 ≥ 100 μg/m 3 , averaged over the period from 1999 - 2011. Fig. 2. The long-term tendency of desert dust over Tel-Aviv based on annual numbers of days with PM 10 ≥ 100 μg/m 3 , during the period from 1999 2011. Fig. 3 and 4. Comparisons between model-predicted surface dust concentrations and PM 10 data taken at the Tel-Aviv University site for two dust events, February 23 – 26, 2007, and March 10 – 14, 2007. Fig. 5. Comparison between model-predicted surface dust concentrations and PM 10 data taken at the Tel-Aviv University site for all consecutive days from February to April 2006. Pavel Kishcha, Desert dust lab work

Attention! Two important comments. Comment 1. To make model-vs-. measurement comparisons (for February 23 – 26, 2007, March 10 – 14, 2007, and for all consecutive days from February to April 2006), you should use model-predicted data from the previous day (18 – 36 hour forecasts) and not from the current day. (See above the example of model-vs. -measurement comparison for March 13, 2006). From the model output file for the previous day, you take numbers from four to seven (X 4, X 5, X 6, and X 7), which correspond to 18 hour forecast, 24 -hour forecast, 30 -hour forecast, 36 -hour forecast respectively. These model forecasts from the previous day you should compare with measured PM 10 concentrations on the current day. Pavel Kishcha, Desert dust lab work

Comment 2. For model-vs. -measurement comparison, you will compare two data sets in Universal Time (UT): the first data set includes PM 10 measurements and the other data set includes model data. First, download hourly PM 10 from the site of the Ministry of Environment Protection. Than, create the first data set by selecting PM 10 data for 2, 8, 14, and 20 LT (local Israel time), which corresponds to 0, 6, 12, and 18 UT (Universal Time). Finally, create the data set of model-predicted dust concentrations. Model data are in UT time. Use model data from the previous day where number 3 corresponds to 0 UT, number 4 corresponds to 6 UT, number 5 corresponds to 12 UT, number 6 corresponds to 18 UT.