The ocean plays a major role in the global redistribution of heat within the climate system, however direct observations of temperature variability in the deep ocean (i.e., below 2000 dbar) are very sparse in the present Global Ocean Observing System. A recent study (Meinen et al., 2020) led by scientists at NOAA/AOML, which garnered interest in mainstream media including Smithsonian Magazine, The Guardian, and more, has demonstrated that an innovative use of an internal temperature sensor located within the sphere of a pressure-equipped inverted echo sounder (PIES) mooring can provide high temporal-resolution (hourly) near-bottom temperature measurements over extended time periods. Two long-term (10-15+ year) arrays of PIES moorings have been maintained in the North and South Atlantic by NOAA and international partners. These long-term records represent ideal data sets to both characterize deep temperature variations with the existing observing system and to evaluate the realism of deep temperature variability in the present generation of numerical ocean models. Furthermore, these data can be used to quantify deep ocean warming in a manner never before accomplished on a broad scale, and can be used to investigate the mechanisms involved in the observed deep/abyssal ocean temperature variability, as well as their implications. For this proposal we will primarily focus on bottom temperature observations collected from the 34.5°S and 26.5°N trans-basin arrays, however as time permits, additional existing records at other latitudes within the Atlantic sector will also be obtained and analyzed. This innovative analysis will aid in quantifying observational uncertainties in the existing long-term observations of deep ocean temperature, as well as help inform decisions about where future observations of temperature are needed within this poorly observed portion of the ocean. The deep temperature records will also serve as an example/reference data set for interpreting future deep temperature variations as new deep observational platforms are brought online (e.g., global deep Argo). As a second component of this proposed work, outputs from a series of different numerical model simulations, state estimates, and operational analyses (e.g., GOFS3.1, OFES, ECCO2, and GFDL simulations) will be examined. The highly-temporally- resolved temperature records from the PIES represent a novel benchmark against which to evaluate the skill of the deep and abyssal temperature evolution in these models. We will use this data to assess the fidelity of the model mean temperatures, their variability, and their trends. Most importantly, once validated, these models will provide the necessary tools to determine the mechanisms driving deep temperature variability at key locations in the Atlantic Ocean.