1: \begin{abstract}

2: {\noindent 

3: We develop a bottom-up approach to flavour models which combine modular symmetry with orbifold constructions.

4: We first consider a 6d orbifold $\mathbb{T}^2/\mathbb{Z}_N$, with a single torus defined by one complex coordinate 

5: $z$ and a single modulus field $\tau$, playing the role of a flavon transforming under a finite modular symmetry.

6: We then consider 10d orbifolds with three factorizable tori, each defined by one complex coordinate $z_i$ and involving the three moduli fields $\tau_1, \tau_2, \tau_3$ transforming under three finite modular groups.

7: Assuming supersymmetry, consistent with the holomorphicity requirement, 

8: we consider all 10d orbifolds of the form $(\mathbb{T}^2)^3/(\mathbb{Z}_N\times\mathbb{Z}_M)$,

9: and list those which have fixed values of the moduli fields (up to an integer).

10: The key advantage of such 10d orbifold models over 4d models is that the values of the moduli are not completely free but are 

11: constrained by geometry and symmetry.

12: To illustrate the approach we discuss a 10d modular seesaw model with $S_4^3$ modular symmetry 

13: based on $(\mathbb{T}^2)^3/(\mathbb{Z}_4\times\mathbb{Z}_2)$

14: where $\tau_1=i,\ \tau_2=i+2$ are 

15: constrained by the orbifold, while $\tau_3=\omega$ is determined by imposing a further remnant $S_4$ flavour symmetry,

16: leading to a highly predictive example in the class CSD$(n)$ with $n=1-\sqrt{6}$.

17: % predicting the atmospheric angle $\theta_{23}$ to lie in the second octant.

18: %Alternatively the remnant $S_4$ flavour symmetry can combine with $S_4^2$ modular symmetry to 

19: %act as an eclectic symmetry to control the K\"ahler potential.

20: }

21: \end{abstract}

22: