On decay–surge population models