According to the neutrino flavour oscillation formula $$ P_{\alpha\beta}=4\sum\limits_{i<j}U_{\alpha i}U_{\beta i}^*U_{\alpha j}U_{\beta j}^* \sin^2\frac{(m_i^2-m_j^2)L}{4E}.$$ Hence, there can be three probabilities on the RHS $P_{e\mu},P_{e,\tau},P_{\mu\tau}$. Let us consider $$P_{e\mu}=4\sum\limits_{i<j}U_{e i}U_{\mu i}^*U_{e j}U_{\mu j}^* \sin^2\frac{(m_i^2-m_j^2)L}{4E}.\tag{1}$$ Expanding (1), we get, $$P_{e\mu}=4U_{e 1}U_{\mu 1}^*U_{e 2}U_{\mu 2}^* \sin^2\frac{(m_1^2-m_2^2)L}{4E}+4U_{e 1}U_{\mu 1}^*U_{e 3}U_{\mu 3}^* \sin^2\frac{(m_1^2-m_3^2)L}{4E}+4U_{e 2}U_{\mu 2}^*U_{e 3}U_{\mu 3}^* \sin^2\frac{(m_2^2-m_3^2)L}{4E}.$$ Similar expressions hold for $P_{\mu\tau}, P_{e\tau}$.
From the Solar neutrino experiment, can/does one measure all probabilities $P_{e\mu},P_{\mu\tau},P_{e\tau}$?
It appears to me that there are various unknowns such as 3 mass squared differences, all 3 mixing angles $\theta_{12},\theta_{23},\theta_{13}$, CP phase. Therefore, how does one measure the solar neutrino mixing angle $\theta_{12}$? Is it possible to eliminate all this unknowns except $\theta_{12}$ in favour of various oscillation probabilities $P_{\alpha\beta}$, $L$, and $E$?
Historically, solar mixing angle was measured first. Am I correct? Without any information of other mixing angles. Therefore, one must have eliminated other mixing angles and phases in favour of $P_{\alpha\beta}$, $L$, and $E$. Is that right?
I'm not interested in experimental subtleties (such as neutrinos from Sun can't be strictly monochromatic and there must be an energy spread etc).