Current General Circulation Models (GCMs) are struggling to capture the observed diurnal cycle of precipitation. One major problem is that GCMs usually have difficulties to simulate nocturnal deep convections. Nocturnal deep convections are usually decoupled with the land surface and are controlled by other factors aloft, which have not been fully understood. The Plains Elevated Convection at Night (PECAN) is a multi-institutional collaborated field experiment conducted at the central Great Plains during summertime (1 June to 15 July) 2015 on the purpose of improving the understanding and simulation of the processes that initial and maintain deep convections at night. During the 45-day PECAN period there are 31 intensive operational periods (IOP) of nighttime deep convections, each last for one night. The unique part of PECAN experiment is the use of 6 fixed and 4 mobile PECAN Integrated Sounding Arrays (PISA) to take detailed measurements for each IOP. These detailed measurements of thermodynamics, wind and water vapor profiles in high temporal resolution (typically 90-min) allow the observation of the evolution of boundary layer and propagating features. We will report results from the analysis of these PISAs using a constrained variational analysis method to understand the large-scale environment that control the nighttime deep convections at the central Great Plains.