The climate research community uses atmospheric reanalysis data sets to understand a wide range of processes and variability in the atmosphere, yet different reanalyses may give very different results for the same diagnostics. The Stratosphere–troposphere Processes And their Role in Climate (SPARC) Reanalysis Intercomparison Project (S-RIP) is a coordinated activity to compare reanalysis data sets using a variety of key diagnostics. The objectives of this project are to identify differences among reanalyses and understand their underlying causes, to provide guidance on appropriate usage of various reanalysis products in scientific studies, particularly those of relevance to SPARC, and to contribute to future improvements in the reanalysis products by establishing collaborative links between reanalysis centres and data users. The project focuses predominantly on differences among reanalyses, although studies that include operational analyses and studies comparing reanalyses with observations are also included when appropriate. The emphasis is on diagnostics of the upper troposphere, stratosphere, and lower mesosphere. This paper summarizes the motivation and goals of the S-RIP activity and extensively reviews key technical aspects of the reanalysis data sets that are the focus of this activity. The special issue The SPARC Reanalysis Intercomparison Project (S-RIP) in this journal serves to collect research with relevance to the S-RIP in preparation for the publication of the planned two (interim and full) S-RIP reports.

Accurate depictions of the tropopause and its changes are important for studies on stratosphere–troposphere exchange and climate change. Here, the fidelity of primary lapse-rate tropopause altitudes and double tropopause frequencies in four modern reanalyses (ERA-Interim, JRA-55, MERRA-2, and CFSR) is examined using global radiosonde observations. In addition, long-term trends (1981–2015) in these tropopause properties are diagnosed in both the reanalyses and radiosondes. It is found that reanalyses reproduce observed tropopause altitudes with little bias (typically less than ±150 m) and error comparable to the model vertical resolution. All reanalyses underestimate the double tropopause frequency (up to 30 % lower than observed), with the largest biases found in JRA-55 and the smallest in CFSR. The underestimates in double tropopause frequency are primarily attributable to the coarse vertical resolution of the reanalyses. Significant increasing trends in both tropopause altitude (40–120 m per decade) and double tropopause frequency (≥3 % per decade) were found in both the radiosonde observations and the reanalyses over the 35-year analysis period (1981–2015). ERA-Interim, JRA-55, and MERRA-2 broadly reproduce the patterns and signs of observed significant trends, while CFSR is inconsistent with the remaining datasets. Trends were diagnosed in both the native Eulerian coordinate system of the reanalyses (fixed longitude and latitude) and in a coordinate system where latitude is defined relative to the mean latitude of the tropopause break (the discontinuity in tropopause altitude between the tropics and extratropics) in each hemisphere. The coordinate relative to the tropopause break facilitates the evaluation of tropopause behavior within the tropical and extratropical reservoirs and revealed significant differences in trend estimates compared to the traditional Eulerian analysis. Notably, increasing tropopause altitude trends were found to be of greater magnitude in coordinates relative to the tropopause break, and increasing double tropopause frequency trends were found to occur primarily poleward of the tropopause break in each hemisphere.

We present in situ observations of convectively injected water vapor in the lower stratosphere from instruments aboard two aircraft operated during the Deep Convective Clouds and Chemistry experiment. Water vapor mixing ratios in the injected air are observed to be 60–225 ppmv at altitudes 1–2 km above the tropopause (350–370 K potential temperature), well above observed background mixing ratios of 5–10 ppmv in the lower stratosphere. Radar observations of the responsible convective systems show deep overshooting at altitudes up to 4 km above the lapse rate tropopause and above the flight ceilings of the aircraft. Backward trajectories from the in situ observations show that convectively injected water vapor is observed from three distinct types of systems: isolated convection, a convective line, and a leading line-trailing stratiform mesoscale convective system. Significant transport of additional tropospheric or boundary layer trace gases is observed only for the leading line-trailing stratiform case. In addition, all observations of convective injection are found to occur within large-scale double-tropopause events from poleward Rossby wave breaking. Based on this relationship, we present a hypothesis on the role of the large-scale lower stratosphere during convective overshooting. In particular, the reduced lower stratosphere stability associated with double-tropopause environments may facilitate deeper levels of overshooting and convective injection.

On 27 August 2013, during the Studies of Emissions and Atmospheric Composition, Clouds and Climate Coupling by Regional Surveys field mission, NASA's ER‐2 research aircraft encountered a region of enhanced water vapor, extending over a depth of approximately 2 km and a minimum areal extent of 20,000 km² in the stratosphere (375 K to 415 K potential temperature), south of the Great Lakes (42°N, 90°W). Water vapor mixing ratios in this plume, measured by the Harvard Water Vapor instrument, constitute the highest values recorded in situ at these potential temperatures and latitudes. An analysis of geostationary satellite imagery in combination with trajectory calculations links this water vapor enhancement to its source, a deep tropopause‐penetrating convective storm system that developed over Minnesota 20 h prior to the aircraft plume encounter. High resolution, ground‐based radar data reveal that this system was composed of multiple individual storms, each with convective turrets that extended to a maximum of ~4 km above the tropopause level for several hours. In situ water vapor data show that this storm system irreversibly delivered between 6.6 kt and 13.5 kt of water to the stratosphere. This constitutes a 20–25% increase in water vapor abundance in a column extending from 115 hP to 70 hPa over the plume area. Both in situ and satellite climatologies show a high frequency of localized water vapor enhancements over the central U.S. in summer, suggesting that deep convection can contribute to the stratospheric water budget over this region and season.

Expansive cirrus clouds present above the anvils of extratropical convection have been observed in satellite and aircraft-based imagery for several decades. Despite knowledge of their occurrence, the precise mechanisms and atmospheric conditions leading to their formation and maintenance are not entirely known. Here, the formation of these cirrus “plumes” is examined using a combination of satellite imagery, four-dimensional ground-based radar observations, assimilated atmospheric states from a state-of-the-art reanalysis, and idealized numerical simulations with explicitly resolved convection. Using data from 20 recent events (2013–present), it is found that convective cores of storms with above-anvil cirrus plumes reach altitudes 1–6 km above the tropopause. Thus, it is likely that these clouds represent the injection of cloud material into the lower stratosphere. Comparison of storms with above-anvil cirrus plumes and observed tropopause-penetrating convection without plumes reveals an association with large vector differences between the motion of a storm and the environmental wind in the upper troposphere and lower stratosphere (UTLS), suggesting that gravity wave breaking and/or stretching of the tropopause-penetrating cloud are/is more prevalent in plume-producing storms. A weak relationship is found between plume occurrence and the stability of the lower stratosphere (or tropopause structure), and no relationship is found with the duration of stratospheric penetration or stratospheric humidity. Idealized model simulations of tropopause-penetrating convection with small and large magnitudes of storm-relative wind in the UTLS are found to reproduce the observationally established storm-relative wind relationship and show that frequent gravity wave breaking is the primary mechanism responsible for plume formation.