Instead of processing them one by one, as suggested before, to achieve better performance you can "steal" underlying containers from queues, merge them and construct another priority_queue from the merged container:
// https://stackoverflow.com/a/1385520/8414561
template <class T, class S, class C>
S& Container(std::priority_queue<T, S, C>& q) {
    struct HackedQueue : private std::priority_queue<T, S, C> {
        static S& Container(std::priority_queue<T, S, C>& q) {
            return q.*&HackedQueue::c;
        }
    };
    return HackedQueue::Container(q);
}
int main()
{
    std::priority_queue<int> queue1{std::less<int>{}, {1,2,3,4,5,6,7}};
    std::priority_queue<int> queue2{std::less<int>{}, {10,14,15}};
    auto v1 = std::move(Container(queue1));
    auto v2 = std::move(Container(queue2));
    v1.insert(v1.end(), std::make_move_iterator(v2.begin()), std::make_move_iterator(v2.end()));
    std::priority_queue<int>queue3{std::less<int>{}, std::move(v1)};
    while (!queue3.empty()) {
        std::cout << queue3.top() << std::endl;
        queue3.pop();
    }
}
Live example.