Questions/Comments • Re: Sunglint calculation

Looks like the calculation of alpha in getglint.f is unnecessary with the assumptions made.
When sigc = sigu,
expon = -(swig**2+eta**2)/2
= -((sin(alphap)*tan(beta)/sigc)**2+(cos(alphap)*tan(beta)/sigu)**2)/2
= -tan(beta)**2/sigc**2 * ((sin(alphap)**2+cos(alphap)**2)/2
= -tan(beta)**2/sigc**2 /2
alphap got canceled out because the sum of its sine squared and cosine squared is always 1.

Then the calculation of slope probability distribution and normalized glint radiance can be made without any wind direction information, which is not needed as one of the inputs with the assumptions made in the simplification.

Relevant codes in getglint.f:
73 sigc = .04964*sqrt(y4)
74 sigu = .04964*sqrt(y4)
75 c endif
76 chi = x5
77 alphap = alpha+chi
78 swig = sin(alphap)*tan(beta)/sigc
79 eta = cos(alphap)*tan(beta)/sigu
80 expon = -(swig**2+eta**2)/2.
81 if (expon.lt.-30.) expon = -30. ! trap underflow
82 if (expon.gt.+30.) expon = +30. ! trap overflow
83 prob = exp(expon)/(2.*pi*sigu*sigc)


Another clarification question:
These two lines are inconsistent.
25 C cos(OMEGA) = cos(BETA)cos(PHI)-sin(BETA)sin(PHI)cos(ALPHA)
59 omega = acoss(cos(y1)*cos(y2)-sin(y1)*sin(y2)*cos(y3))/2.

I assume the code (line 59) is correct, and the correct formula on line 25 should be cos(2*OMEGA) = cos(BETA)cos(PHI)-sin(BETA)sin(PHI)cos(ALPHA)?

Eq. 3 in Cox and Munk (1954) is consistent with line 25 of getglint.f, also missing a factor of 2 in front of omega?

Statistics: Posted by starbright — Fri May 10, 2024 12:54 pm America/New_York

---