Characterization of the candidate site for the Cherenkov Telescope Array at the Observatorio del Teide
Irene Puerto-Giménez, Markus Gaug, Rafael Barrena, Julio Castro, Michele Doro, Lluís Font, Miguel Nievas Rosillo, Jaime Zamorano
aa r X i v : . [ a s t r o - ph . I M ] J u l RD I NTERNATIONAL C OSMIC R AY C ONFERENCE , R
IO DE J ANEIRO T HE A STROPARTICLE P HYSICS C ONFERENCE
Characterization of the candidate site for the Cherenkov Telescope Array at theObservatorio del Teide
I. P
UERTO -G IM ´ ENEZ , M. G AUG , , R. B ARRENA , J. C ASTRO , M. D ORO , , , L L . F ONT , , M. N IEVAS R OSILLO , J. Z AMORANO , FOR THE
CTA C
ONSORTIUM . Instituto de Astrof´ısica de Canarias, 38205 La Laguna, Tenerife, Spain F´ısica de les Radiacions, Departament de F´ısica, Universitat Aut`onoma de Barcelona, 08193 Bellaterra, Spain. CERES, Universitat Aut`onoma de Barcelona-IEEC, 08193 Bellaterra, Spain. University and INFN Padova, Via Marzolo 8, 35131 Padova, Italy. Depto. F´ısica At´omica, Universidad Complutense de Madrid, 28040 Madrid, Spain Depto. Astrof´ısica y Ciencias de la Atm´osfera, Universidad Complutense, 28040 Madrid, Spain [email protected]
Abstract:
The Spanish partners of the future Cherenkov Telescope Array (CTA) have selected a candidate sitefor the Northern installation of CTA, at 3 km from the Observatorio del Teide (OT), in the Canary Island ofTenerife. As the OT is a very well-characterized astronomical site. We focus here on differences between thepublicly accessible measurements from the OT observatory and those obtained with instruments deployed at thecandidate site. We find that the winds are generally softer at the candidate site, and the level of background lightcomparable to the Observatorio del Roque de los Muchachos (ORM) at La Palma in the B-band, while it is onlyslightly higher in the V-band.
Keywords:
Observatorio del Teide, CTA, Instrumentation and Methods for Astrophysics, IACT
The Canaries have two islands with world-class skies forastronomy, which can compete for hosting CTA-North.These islands host two international observatories whichtogether constitute the most important optical, infraredand gamma-ray observatories in Europe: the Observatori-o del Teide (OT) at the island Tenerife, and the Observa-torio del Roque de los Muchachos (ORM) at the islandLa Palma. A specific National Law protects the qualityof the sky in these observatories [1]. As La Palma doesnot fulfill the current CTA requirement of having flat landof 1 km above 1,500 meters, we have selected a site inTenerife, close to the OT. This site is a plateau at an al-titude of 2,260 m a.s.l. with the geographic coordinates28 ◦ ◦ Temperature measurements.
The average temperature at
Fig. 1 : The evolution of the average annual temperature,average annual minima and maxima at IZO [3].the OT is 9.8 ◦ C, with minimum and maximum averagesbeing 5.9 ◦ C and 13.6 ◦ C respectively [2]. These agree withthe temperatures measured at IZO, yielding an average an-nual temperature around 10 ◦ C, for the period from 1915to 2000, and average minima and maxima around 6 ◦ C and13 ◦ C, respectively (Fig. 1).
Wind Measurements.
Due to the abrupt orography of theisland, wind speed and direction greatly vary from place toplace. Wind speed data at IZO are available since 1933 andare publicly accesible from the AEMET webpage [5]. Thewind speed measurements were taken every second or ev-ery 0.25 seconds (depending on the anemometer which haschanged over the years). The average wind speed valueshave been averaged every 10 minutes. The IZO anemome-ter was placed at different heights: from Jan 1933 to Feb1984 at 12 m height, then until Feb 2000 at 16 m and thento present at 10 m height. Fig. 2 presents the monthly av-erage wind speed cumulative frequency at IZO from 1933to present: For over 90% of the values, the average windspeed is lower than 9 m/s (32.4 km/h). The average valuelies at 6.8 m/s (24.6 km/h). It is however expected that thewind speeds are lower at the CTA candidate site (CTA-s), haracterization candidate site for CTA at the OT33 RD I NTERNATIONAL C OSMIC R AY C ONFERENCE , R
IO DE J ANEIRO since IZO is located at the top of the most exposed peak inthe area, whereas the CTA-s is more shielded.
Fig. 2 : Cumulative monthly average wind speed frequency,as observed at IZO [4] from Jan 1933 to Aug 2011.
Fig. 3 : Simulatenous wind speed (left) and wind direction(right) measurements at IZO and CTA-s, taken betweenAug 2011 and Apr 2012.In Fig. 3, we show a comparison of simultaneous windmeasurements at IZO and the CTA-s, from Aug 2011 toApr 2012. We compare wind at 10 m height at IZO withwind at 2.5 m height at the CTA-site (Fig. 3-left) andfind that the wind speed at the CTA-s is always about afactor 3 lower than at IZO. In Fig. 3-right we compare thesimultaneous wind direction, and find again that there is aquite strong global correlation between the two sites.To test whether the difference in the position of theanemometers can account for the large difference in windspeed observed at CTA-s compared to IZO, we comparemeasurements taken both at CTA-s with the sensor at 2.5 mand 10 m above the ground, from Dec 2012 to April 2013.Fig. 4 shows the results. We see a clear correlation, androughly an increase of a factor 2 in wind speed from 2.5 mto 10 m. Compared to Fig. 3 we still observe that the typi-cal CTA-s wind speed is about two thirds of that at IZO.
Humidity Measurements.
The values presented here cor-respond to measurements taken at IZO and at the OT Themedian, mean and percentiles (5%, 25%, 75% and 95%)are given in Fig. 5. The median is found at 29% humidity,and in 75% of the night-time, the relative humidity is low-er than 55%. The median humidity value of 29%, foundfor this one year of data, coincides with the value obtainedfrom a 5-year site characterization campaign for the Eu-ropean Solar Telescope (EST), carried out from 2003 to ] (wind station @2.5m) -1 [km.h 〉 wind v 〈 -5 0 5 10 15 20 25 30 35 ] ( w i nd s t a t i on @ m ) - [ k m . h 〉 w i nd v 〈 Entries 6609Mean x 7.381Mean y 16.41RMS x 4.436RMS y 9.369Entries 6609Mean x 7.381Mean y 16.41RMS x 4.436RMS y 9.369
Fig. 4 : Comparison of wind speeds measurements at CTA-s taken with the anometer placed at 2.5 m and 10 m abovethe ground. Courtesy St. Vincent.
Fig. 5 : Frequency of daytime (top) and night-time (bottom)humidity, as observed at IZO during 10 years. In blue,the normalized frequency while in black the cumulativedistribution is shown. The sensor used for this campaignsaturates at values around 99%, which causes the abruptchange for the highest humidity values. These correspondto days with overspill of clouds, where the observatory isinside the clouds.2008 at the OT. However, the highest humidity values areless frequent in this long database [6].
Precipitation Measurements.
The climate at the IZO re-gion is extremely dry most of the year, as it lies above thequasi-permanent temperature inversion layer, in the dryfree troposphere. Precipitation is mainly recorded in win-ter time when Atlantic low areas pass over the Canary Is-lands. Fig. 6 presents the seasonal frequency of days ofappreciable precipitation (i.e. precipitation > haracterization candidate site for CTA at the OT33 RD I NTERNATIONAL C OSMIC R AY C ONFERENCE , R
IO DE J ANEIRO
Fig. 6 : Frequency of precipitation-days per month, mea-sured at IZO and averaged over 29 years [7]. &
10 mm, the annual number of days goes down to 11 pre-cipitation days per year. The AEMET public database alsogives the number of days with hail per month from 1920to 2011. The annual average for this period amounts to 1.5days of hail per year.
The Cherenkov light that arrives to the telescopes fromthe particle showers in the high atmosphere is highly de-pendent on the quality of the atmosphere that it travers-es. In fact different density profiles lead to differences inCherenkov light density of up to 60% [8], and the pres-ence and position of clouds and aerosol layers also affectdifferently the atmosphere transmission [9].
Cloud Coverage and Useful Nights.
The atmosphere inthe subtropical region of the Canary Islands is character-ized by its great stability throughout the year. This is dueto the combination of two processes of the atmospheric cir-culation at large scale [10]. One of them is the descend-ing branch of the Hadley cell around 30 ◦ N and the otherone the Trade or Alisios Winds coming from the Azoreshigh area that blow in the low troposphere above an oceanwhich is relatively cold. As a result, a temperature inver-sion layer appears around 1,300 m a.s.l. on average, calledthe “Alisio inversion”, which can be usually well identifiedby the sea of stratocumulus on the Northern coasts of theislands. This layer separates two well-defined regimes: be-low it, there is the moist marine boundary layer and aboveit, the dry free troposphere (where the OT and the CTA-sare located). The Alisio inversion is a quasi-permanent lay-er, being present 78% of the time throughout the year. It-s altitude and thickness has a seasonal dependence, beinghigher and thinner during the winter (when it is located be-tween 1,350 and 1,850 m a.s.l., being only 350 m thick)and lower and thicker during the summer between 750 and1,400 m a.s.l., being about 550 m thick) [11]. Unfortunate-ly, so far there are no specific and well-calibrated instru-ments for ground measurements of cloud coverage at theOT or IZO. We present here a compendium with the mostrelevant data found: sunshine measurements at IZO anduseful time at the ORM observatory on La Palma, 140 kmaway and also located at 2400 m a.s.l., i.e. also above theinversion layer, and cloud coverage at the ORM calculatedwith satellites.The insolation, defined as the period of time duringwhich the sun shines, at IZO is very high, especially insummer, recording the highest average annual insolationduration of Spain with 3448.5 hr/year. Fig. 7 shows the
Fig. 7 : Monthly insolation, as observed at IZO [12].monthly insolation in percentage taken with a heliographat IZO from Jan 2001 to Dec 2011. The average is 83% ofsunshine. Detailed satellite studies have been carried outby the certified consulting meteorologist Dr. Andr´e Eras-mus for the ORM on La Palma, showing that the photo-metric time at this observatory is 83.7% (for the 7-year pe-riod 1996-2002). Cross-calibration with ground measure-ments using the Carlsberg Meridian Telescope (CMT) atthe ORM shows a good agreement of the data [13], withonly 1.2% differences [14].The weather downtime calculated [15] for the period1999-2003 using the logs of the CMT telescope obtainedan average of 20.7% weather downtime (defined as theperiod when there were no recorded observations during awhole night). The monthly average of weather downtimewas also calculated and compared with the one obtainedat the William Herschel telescope (WHT) for the 18-yearperiod from 1989 to 2006 and it is in good agreement withthe previous numbers, see Fig. 8.
Fig. 8 : Relative monthly frequency of weather downtimeat the CMT, averaged over 5 years, and compared with thatof the WHT for an 18-year period [16].Other in-situ measurements of useful time at the ORMyield the values: 78% [17], 75% [18] (11-year time inter-val) and 72.7%-77.5% [19] (4-year period). More detailedreferences and data can be found at the IAC Sky QualityGroup webpage [20]. At both observatories, the percent-age of time lost per year due to adverse atmospheric con-ditions remains between 16.3% and 27.5%, depending onthe methodology for calculations and the selected periodof time.
In-site Particle Matter.
Cuevas et al. [21] concludes thatin-situ TSP, PM10 data, and Aerosol Optical Depth (AOD)observations performed at IZO demonstrate that the siteIZO-OT is characterized by extremely clean air and pris- haracterization candidate site for CTA at the OT33 RD I NTERNATIONAL C OSMIC R AY C ONFERENCE , R
IO DE J ANEIRO tine skies. IZO-OT is sometimes affected by dust-loadedAfrican air mass intrusions in the summer time (Jul-Sep),called calima . Since African air masses impact only insummer when nights are short, the annual percentage ofnocturnal observing affected by calima is very low. In ad-dition, most of the dust uses to keep below the inversionlayer and hardly affect the observatory. Finally, the factthat under calima intrusions, middle clouds are normallyobserved between 5 and 6 km altitude minimizes the neg-ative impact of these intrusions on astronomical observa-tions performed at the OT. Overall, the low impact of cali-ma events can be seen in Fig. 7 where the weather down-time percentage in summer months is lower than 10% atthe ORM.
Night Sky Background (NSB).
Here we present the re-sults of a document about NSB at OT and ORM by M.Nievas-Rosillo, supervised by R. Barrena-Delgado, fromsummer 2012, to be published soon. The following tableshows NSB measurements in the V-band (mag/arcsec ) forthe OT taken with 3 different instruments, an all-sky cam-era of type AstMon, an 80 cm telescope (the IAC80) andthe sky-quality-meter from Unihedron (SQM), similar tothe one used in the CTA-ATMOSCOPE, corrected for thelight contamination by the Milky Way:Instrument All Data From 2011AstMon (From 2012) 21,41 ± ± ± ± ± ± ) [22] are compared between OT and ORM(both are telescope measurements) in the next table:Instrument B VIAC-OT 22,34 ± ± ± ± ± ± , ± ,
24 mag/arcec . The site proposed by the Spanish partners of CTA at 3 kmfrom the OT consitutes an excelent candidate to host CTA-North. Actually, OT, together with ORM in La Palma, is avery extensively and long-term characterized astronomicalobservatory. Such a long-term database makes the knowl-edge of the atmospheric conditions very reliable and re-duces the risk of a wrong characterization from short-termmeasurements. We have summarized the main climate andatmospheric parameters obtained from public databases. Inorder to complement the public databases, some measure-ments have been taken with instruments deployed at thecandidate site. Of special importance is the result we haveobtained by comparing wind speeds: the wind speed at thecandidate site is about one thirds lower than that at IZO, ac-cording to the fact that the site is more shielded than the I-ZO anemometer. Recent results on the NSB measurements at the OT compared with ORM show that even though thenight sky at OT is brighter than that of ORM, the differ-ence in the B band is very low and will hardly affect theCTA performance.
Acknowledgements:
We gratefully acknowledge sup-port from the agencies and organizations listed in thispage:
References [1] Ley 31/1988 del 31 de Octubre.[2] See [3] Graph from IARC webpage on climatology [4] Data from AEMET database [5] [6] Sky Quality Group of the IAC, private communication.[7] Graph from IARC webpage on climatology [8] K. Bernl¨ohr, Astropart.Phys. 12 (2000) 255-268[9] D. Garrido et al, these proceedings ID 465[10] see e.g. Palm´en E. and Newton C.W, InternationalGeographic ser. 13, Academic Press Ed (1969)[11] see e.g. Torres C. et al., 3 rd Hispanic-PortugueseAssembly of Geodesics and Geophysics (2002).[12] Data from AEMET database [13] Erasmus, D.A. and van Rooyen, R. Final Report to ESO2006, Purchase Order 73526/TSD/04/6179/GW/LET[14] Mu˜noz-Tu˜n´on, C., Varela, A. M. and Fuensalida, J. J.RevMexAA 31 (2007) 36-46.[15] Garc´ıa-Gil, A., Mu˜noz-Tu˜n´on, C., and Varela, A. M.,PASP 122:1109-1121, 2010.[16] Graph from [15].[17] Murdin P., Vistas in Astronomy 28, Issue 2 (1985)449-465.[18] ING Annual Report 1999 of PPARC-NWO ING Board,and Rutten, R.G.M “ING Director’s Report: operationsand development”, 9-10 April 2001.[19] Lombardi, G., Zitelli, V., Ortolani, S., Pedani, M., PASP,118 (2006) 1198-1204[20]