1. MODIS Level 3: Future Issues and Considerations Brent Maddux MODIS Level 3: Future Issues and Considerations Brent Maddux 2001-2004 Cloud Fraction Mean 1 0.5 0

2.  Do they represent known physical phenomenon well?  What is L3 saying? ₪ Some limitations of L3 SDSs ₪ How do we interpret the data effectively?  decrease the limitations  improve physical meaning and interpretation How do the MODIS L3 products look: Going from information to knowledge: Next steps:

3. DJF High Cloud DJF All Cloud HIRS Aqua Terra Aqua 1.0 0.8 0.6 0.4 0.2 0.0

4. Dust or high aerosol events 200 400 600 800 1000 mb Cloud Top Pressure May 1 st, 2003 ₪ Events last several days ₪ Few if any visible clouds ₪ 750-900 mb ₪ For daily data this can be accounted for, but not weekly or monthly.

5. 0 200 400 600 800 Aqua Terra Difference -800 -400 0 400 800 Cloud Ice Water Path December 2002 Major ice cloud events? g/m^2

6. 0 150 300 Difference Cloud Liquid Water Path December 2002 Aqua Terra g/m^2 -150 0 150 Major liquid cloud events?

7. Possible Solutions: 0 15 30 Ocean Mass Concentration ₪ Conduct secondary data processing ₪ Remove questionable data ₪ Create uncertainty products 0 500 1000 Cloud Top Pressure 1x10^-6 g/m^2mb

8. 0 15 30 45 60 µm 0 200 400 600 800 0 250 500 750 1000 g/m^2 mb Ice Particle Size and Tropical Deep Convection Evolution -Uniform re ice over cores -Isolated deep convection -Evident cirrus shields ₪ Ice particle size increase away from deep convection Ice Particle Radius Ice Water Path Cloud Top Pressure ₪ What would scatter plots show?

9. Mean SDS Comparisons ₪ Mean phase properties can’t be fully characterized

10. 0 15 30 45 60 Ice Effective Radius (µm) Cloud Top Pressure (mb) Ice Effective Radius (µm) Water Path of Ice (g/m^2) 0 1000 2000 0 500 1000 IWP vs Re Ice CTP vs Re Ice ₪ cloud regimes ₪ cloud height or phase ₪ particle size or path Ambiguous Property comparison limitations: ₪ data subsets (large particle size, phase, etc) ₪ multiple data set comparison Possible solutions: ₪ new joint histogram (in collection 5) ₪ means of property subsets ₪ multi dimensional histograms Mean Ice Properties:

11. Summary and Conclusions: ₪ Implement a secondary processing between L2 and L3. ₪ Add additional SDSs 1) that can be used to further compare the microphysical properties of clouds and other atmospheric parameters 2) that allow for a further intuitive interpretations 3) near nadir.

12. Image courtesy of MODIS Rapid Response Project at NASA/GSFC

17. Nov 2001

18. October 2002