Mid-infrared spectrometers are now regarded as versatile sensing solutions for an increasing number of applications, such as fugitive gas detection, combustion diagnostics, and urban chemical spill monitoring. This dissertation explores various novel spectrometers for these applications, with emphasis on system integration, fieldability, and other application-specific demands including size, power, and cost constraints. We will discuss methane detection on an integrated silicon photonic waveguide platform, distinction of an important radical species amongst many coexisting species in a combustion environment, and system construction and field deployment of a quantum cascade laser-based dual-comb spectrometer. In particular, we focus on the management of coherent noises that are unpredictable and highly dynamic, which are dominant sources of noise in the above-mentioned systems and cannot be resolved through traditional means. We investigate the varying causes of these noises in the sensing systems, and in each case targeted solutions are proposed and implemented to extend the sensor capability and stability.