1. Group refractive index ● Method ● Additional runs ● Wavelength distribution ● Systematics ● Results

2. Method Optical Beacon run: Flash light from a given position Measure time when light reaches PMT => velocity of light in water and refractive index Advantage of light from Optical Beacons: Small light dispersion, because of small wavelength range (around 10 nm) Light emitted at wavelength (red new data) 385, 400, 440, 460, 470, 505, 518 and 532 nm

4. Method: Times versus distances Fit linear function => v_measured n=v_vacuum/v_measured Each position has up to three measurements with errors in position of 0.2 m and time with 0.2 ns

5. Additional runs Old runs (69 runs taken between 2008 and 2010) : Light emitted at wavelength 400, 470 and 532 nm New runs (8 times 6 runs taken between 2010 and 2011) Light emitted at wavelength 385, 400, 440, 470, 505, 518 nm Laser runs taken at similar period 532 nm Calibration used 2010:V6.5 for new analysis

6. Old measurements Data at 470 nm (42 runs)

7. New Measurements Data at 400 nm (8 runs)Data at 505 nm (8 runs) Similar for other wavelengths

8. Absorption length distribution Use absorption length distribution: 1) to correct for wavelength distribution at given distance 2) uncertainty for systematics Input: Wavelength distribution at emission place (distribution from spectrometer) Take into account absorption Output: Wavelength distribution at detection place (shifted distribution used in public plot)

9. Change of wavelength distribution through absorption 385 nm Laser 532 nm 460 nm Light curve emitted Expected light curve at 120 meters Drastic change Wavelength distribution used for final plot

10. Evolution of wavelength peak values as function of distance Most wavelength with small changes Largest change for 518 nm 518 nm at 0 m 490 nm at 120 m Width of wavelength distribution also change (not shown for clarity) Analysis range between 100 m and 250 m

11. Scattering as function of wavelength Sea campaign of the year 2000 suggests to take for all wavelength a minimum scattering value of 20 m and a maximum scattering value given by the value used in ANTARES Higher limit given by Mobley Lower limit given by Mobley Used in Antares Lower value used for systematics

12. MC checks with three different refraction indexes (no scattering) Method gives reliable results

13. MC with three different refraction indexes (default values) Variation between input refraction index and measured refraction index (< 0.3%) Variation depends more on scattering length than absorption length

14. MC with different scattering lengths 0.2% Shorter scattering length -> larger measured refraction index 0.3% 0.2% Variation

15. Comparison of parametric formula and measurement Errors for refraction index => 0.3% (scattering) Errors for wavelength distribution from absorption corrections

16. Conclusion ● Runs analyzed from 2008 until 2011 ● Absorption changes the measured wavelength distribution => apply corrections for final plot ● Systematics ● Parametric formula and data agrees ● Paper draft soon ready