This paper investigates a location-based downlink transmission scheme to provide diverse services for different users in an integrated satellite-terrestrial network (ISTN). Specifically, the satellite network employs multicast communication to disseminate information to multiple satellite users, while the terrestrial network incorporates non-orthogonal multiple access (NOMA) with intelligent reflecting surface (IRS) technology to serve terrestrial users. Given that the location information-based channel state information (LoI-CSI) of each user is available, we formulate an optimization problem to minimize the outage probability (OP) of the terrestrial network by optimizing the transmit power and beamforming (BF) weight vector at the base station, the IRS phase shift vector, and the power allocation factor, while meeting the quality-of-service (QoS) requirement of the satellite network. To make the optimization problem tractable, we first propose a low-complexity BF algorithm based on the LoI-CSI, which simplifies the optimization problem while guaranteeing the QoS requirement of the satellite network. Then, assuming that terrestrial links experience Rician fading, we derive an approximate yet accurate OP of the terrestrial network, which is explored to calculate the phase shift vector. Furthermore, we propose a novel power allocation method that employs an exponential-type approximation of the first-order Marcum Q-function, to obtain the power allocation coefficient. Finally, simulation results confirm the theoretical formulas' validity and reveal the proposed algorithms' superiority in system performance.