Underwater communication with autonomous underwater vehicles (AUVs) has strong demands on the modems caused by the constantly changing signal propagation (multi-path propagation, scattering, diffraction and refraction at thermal layers, etc.) of the underwater channel. These demands typically lead to a modem designed to match specific conditions. In this paper we present an automated model-based physical layer implementation approach of an underwater acoustic passband modem using QPSK. This paper presents a possibility to automatically simulate, validate and synthesize a physical layer implementation with a significantly improved ease-of-use, maintainability and extensibility of the developed modems. The results are based on a prototype with six variants of adapted Reed Solomon and convolutional en-/decoders, developed and implemented in hardware on an Altera FPGA. In this work we show that the implemented prototype requires up to 88.7 % less energy for the same or better functionality compared to the Micromodem-2. In addition we show that our hardware can be used immediately in the area of long-distance communication based on a SNR test and in areas of high multi-path propagation based on channel simulation using BELLHOP.