We study numerically the dynamical instabilities and splitting of singly and doubly quantized composite vortices in nonrotated two-component Bose-Einstein condensates harmonically confined to quasi two dimensions. In this system, the vortices become pointlike composite defects that can be classified in terms of an integer pair (κ 1 ,κ 2 ) of phase-winding numbers. Our numerical simulations based on zero-temperature mean-field theory reveal several vortex splitting behaviors that stem from the multicomponent nature of the system and do not have direct counterparts in single-component condensates. By calculating the Bogoliubov quasiparticle excitations of stationary axisymmetric composite vortices, we find complex-frequency excitations (dynamical instabilities) for the singly quantized (1,1) and (1,−1) vortices and for all variants of doubly quantized vortices, which we define by max j=1,2 |κ j | = 2. While the predictions of the linear Bogoliubov analysis are confirmed by direct time integration of the Gross-Pitaevskii equations of motion, the latter also reveals intricate long-time decay behavior not captured by the linearized dynamics. Firstly, the (1,±1) vortex is found to be unstable against splitting into a (1,0) and a (0,±1) vortex. Secondly, the (2,1) vortex exhibits a two-step decay process by splitting first into a (2,0) and a (0,1) vortex followed by the off-axis splitting of the (2,0) vortex into two (1,0) vortices. Thirdly, the (2,−2) vortex is observed to split into a (−1,1) vortex, three (1,0) vortices, and three (0,−1) vortices. Each of these exotic splitting modes is the dominant dynamical instability of the respective stationary vortex in a wide range of intercomponent interaction strengths and relative populations of the two condensate components and should be amenable to experimental detection. Our results contribute to a better understanding of vortex physics, hydrodynamic instabilities, and two-dimensional quantum turbulence in multicomponent superfluids.The aforementioned studies of vortex splitting [23][24][25][26][27][28][29][30][31][32][33][34][35][36][37][38] were conducted for a solitary scalar BEC, which is described by a single C-valued order parameter. However, vortex physics becomes much more diverse when multiple, say K ∈ N, scalar condensates come into contact, interact with one another, and thereby constitute an K-component BEC described by a C K -valued vectorial order parameter. Already the simplest multicomponent system, the two-component BEC corresponding to K = 2, has been found to exhibit many stable vortex structures not encountered in single-component BECs, such as coreless vortices [6], square vortex lattices [47][48][49], serpentine vortex sheets [50], triangular lattices of vortex pairs [49], skyrmions [51][52][53], and meron pairs [54,55]. Although presently only the coreless vortices [6] and square vortex lattices [48] from this list have been verified experimentally, studies of exotic vortex configurations in twocomponent BECs are becoming more and more within reach ...