Faced with unprecedented growth in earth data volume and demand, NASA has developed the Earth Data Analytic Services (EDAS) framework, a high performance big data analytics and machine learning framework. This framework enables scientists to execute data processing workflows combining common analysis and forecast operations close to the massive data stores at NASA. The data is accessed in standard (NetCDF, HDF, etc.) formats in a POSIX file system and processed using vetted tools of earth data science, e.g. ESMF, CDAT, NCO, Keras, Tensorflow, etc. EDAS utilizes high performance parallel data access, a custom distributed array framework, and a streaming parallel in-memory workflow for efficiently processing huge datasets within limited memory spaces with interactive response times. EDAS services are accessed via a WPS API being developed in collaboration with the ESGF Compute Working Team to support server-side analytics for ESGF. The API can be accessed using direct web service calls, a Python script, a Unix-like shell client, or a JavaScript-based web application. New analytic operations can be developed in Python, Java, or Scala (with support for other languages planned). Client packages in Python, Java/Scala, or JavaScript contain everything needed to build and submit EDAS requests. The EDAS architecture brings together the tools, data storage, and high-performance computing required for timely analysis of large-scale data sets, where the data resides, to ultimately produce societal benefits. It is currently deployed at NASA in support of the Collaborative REAnalysis Technical Environment (CREATE) project, which centralizes numerous global reanalysis datasets onto a single advanced data analytics platform. This service enables decision makers to compare multiple reanalysis datasets and investigate trends, variability, and anomalies in earth system dynamics around the globe. EDAS services include configurable high performance neural network learning modules designed to operate on the products of EDAS workflows. As a science technology driver we have explored the capabilities of these services for long-range forecasting of the interannual variation of important regional scale seasonal cycles. Neural networks were trained to forecast All-India Summer Monsoon Rainfall (AISMR) one year in advance using (as input) the top 8-64 principal components of the global surface temperature and 200 hPa geopotential height fields from NASA’s MERRA2 and NOAA’s Twentieth Century Reanalyses. The promising results from these investigations illustrate the power of easily accessible machine learning services coupled to huge repositories of earth science data.

Burn severity influences various biophysical and biogeochemical processes, and it is projected to increase in the coming decades across high-latitude ecosystems. However, the impact of burn severity on thermokarst (e.g., land subsidence after ground ice melting) is not well understood. We used time-series image analysis to assess the effects of burn severity and ground ice content on thermokarst processes in the Noatak National Preserve, northwestern Alaska. To extend the temporal depth for illustrating fire-thermokarst linkages, we evaluated eight existing fire indices derived from visible and near-infrared (380-1100 nm) spectral bands and developed burn severity maps of historical fires using Landsat MSS sensors (operation between 1972-1992). Our results reveal that tundra fire is a significant factor in creating thermokarst landforms (p<0.01), and that the magnitude of thermokarst varies with burn severity levels and ground ice content (p<0.05). An abrupt increase in thermokarst occurred one year after fire but the rate of thermokarst decreased after three years. The area of thermokarst three years after fire was highest in high-severity burns (385 ± 47 m2 thermokarst area/ha burned area), followed by moderate- (255 ± 32 m2/ha) and low-severity (201 ± 42 m2/ha) burns. Ground ice content interacted with burn severity to affect thermokarst; the area of thermokarst was twice as large in landscapes with high ground ice (356 ± 67 m2/ha) as in landscapes of low ground ice (167 ± 39 m2/ha) three years after fire. Among the eight fire indices, the Global Environmental Monitoring Index (GEMI) demonstrates the strongest correlation with field-based estimates (R2 = 0.8). Burn-severity maps reconstructed with GEMI reveal that over a 40-year study period, thermokarst expansion occurred more rapidly in high-severe burns than in low-severity or unburned areas. Our results suggest that the projected increase in burn severity may result in abrupt and long-lasting permafrost degradation in tundra ecosystems with potential consequences on Arctic carbon stocks.

Azimuthal anisotropy quantified by teleseismic SKS, SKKS, PKS (“XKS”) and Local S wave splitting parameters is used to investigate mantle deformation and flow beneath the boundary of the North American and Caribbean plates and adjacent areas. A total of 2556 XKS and 299 pairs of local S wave splitting parameters were obtained at 23 stations. The observations can be divided into two groups based on the spatial distribution of the resulting fast polarization orientations. Those observed on the Caribbean Plate are mostly WNW-ESE which are mostly trench-parallel. In contrast, the fast orientations observed on the North American Plate are dominantly NNE-SSW which are approximately trench-orthogonal and are consistent with those previously observed in southern Mexico to the north of the area of the current study. At most of the stations at which XKS and local S wave splitting parameters are available, the splitting parameters from the two types of shear waves are comparable, suggesting that the observed azimuthal anisotropy is mostly from the mantle wedge above the slab. The observations especially those obtained at recently deployed stations in Guatemala provide new insights into the complicated mantle flow system associated with slab subduction and rollback, as well as lithospheric shearing along the southern boundary of the North American Plate.

The Mekong river is one of the most complex river systems in the world that is shared by six nations in Southeast Asia. The river still remains relatively undammed (most existing dams are in the tributaries and are small), and its hydrology today is dominated by large natural flow variations that support the highly productive agricultural and riverine ecological systems; however, this is changing due to the alterations in land use and construction of new dams both in the tributaries the mainstream. Understanding the changes in surface water dynamics is therefore crucial to provide realistic future predictions of changes in downstream floodplain and riverine ecology due to the construction of dams in the upstream. While the existing dams have caused little impact on mainstream flows, those under construction and planned are likely to cause severe and potentially permanent damage to downstream hydro-agro-ecological systems, and adversely impact the livelihood of millions. Here, using hydrodynamic model simulations (CaMa-Flood), we show that the effects of flow regulation on downstream river-floodplain dynamics are relatively predictable along the mainstream Mekong, but flow regulations could potentially disrupt the flood dynamics in the Tonle Sap River (TSR) and small distributaries in the Mekong Delta. Results suggest that TSR flow reversal could cease if the Mekong flood pulse is dampened by 50% and delayed by one-month. While flood occurrence in the vicinity of the Tonle Sap Lake and middle reach of the delta could increase due to enhanced low flow, it could decrease by up to five months in other areas due to dampened high flow, particularly during dry years. Further, areas flooded for less than five months and over six months are likely to be impacted significantly by flow regulations, but those flooded for 5-6 months could be impacted the least.

The key to reproducible data reduction and data processing in scientific research is the ability to faithfully record every step of the process in a reproducible format that is transparent and easy to communicate. This is not an easy task. Most of the time in experimental research that is not primarily computational in nature, it falls victim to the enormous effort required to design experiments well, run complex analytical procedures rigorously and efficiently, and interpret the results in the proper geologic, geochemical or biological context, with little time left to invest in documenting and constructing a reproducible data reduction workflow. While this is understandable, it introduces a high risk for error, makes it extremely difficult to share and discuss one’s approach or review others’, reproduce the calculations at a later point or even just revisit what was done conceptually. Part of the problem lies inherently with most data processing being difficult to document, part of black box process where the inner workings are inaccessible, or simply too divorced from the narrative of the scientific work it represents. One important obstacle that interferes frequently with attempts to remedy this situation in the stable isotope community is the lack of many basic computational and data access tools that enable the kinds of calculations and data processing isotope geochemists need to do on a day to day basis. Here, we introduce a new suite of software packages that provide efficient and transparent access to raw stable isotope ratio mass spectrometry (IRMS) data formats and enable reproducible data processing straight from raw analytical output through data reduction, quality control, visualization and data reporting that retains the necessary flexibility required for the enormous breadth of analytical goals in the stable isotope community. Presented tools will cover aspects of functionality provided by the isoreader (isoreader.kopflab.org), isoviewer (isoviewer.kopflab.org) and isoprocessor (isoprocessor.kopflab.org) software packages and are 100% open-source and freely available to everyone in the geochemical community and beyond.

Louisiana is undergoing rapid change from natural and anthropogenic forces, such as sea level rise, subsidence, and eutrophication. Sediment diversions on the lower Mississippi River are proposed as a large-scale restoration strategy to create new wetlands and sustain existing wetland areas in Barataria Basin, Louisiana. This will introduce a large volume of sediment and nutrient rich freshwater from the Mississippi River to the receiving basins. This will result in, at least, short term changes in light and nutrient dynamics and has potential to alter phytoplankton composition. In order to understand nitrogen dynamics in Barataria Basin due to large scale coastal restoration practices, the nitrogen budgets (including particulate and dissolved forms) were calculated from outputs of the Integrated Biophysical Model, which is based on the existing Delft3D model coupled with a water quality model (D-WAQ). Creating nitrogen budgets in estuarine systems allows for better understanding of the major sources, sinks, inputs and exports across the system, increasing understanding of the amount of nitrogen available to drive estuarine primary production. Quantification of nitrogen inputs, outputs and processes is essential because it is the limiting nutrient for most estuarine primary producers (e.g., phytoplankton and emergent macrophytes). Preliminary model results for the existing conditions suggest that the dissolved inorganic nitrogen in the estuarine waters is mainly derived from diffusional sediment fluxes and mineralization of particulate organic nitrogen. Most of the dissolved inorganic nitrogen was assimilated for phytoplankton growth. A relatively small portion of dissolved inorganic nitrogen was removed from the system through denitrification in the water column. More particulate organic nitrogen originated from emergent macrophytes than from phytoplankton primary production. These model results will help better understand how proposed sediment diversions on the lower Mississippi River may change the future ecological conditions of estuarine open water in coastal Louisiana.

Storm-time geomagnetic disturbances induce significant geoelectric fields within the Earth that can adversely affect the operation of electric power grids. The recently completed magnetotelluric survey supported by the NSF EarthScope program (2006-2018) has produced a large public archive of impedance tensors across much of the continental United States (US). In this work, the EarthScope tensors are convolved with long time series of geomagnetic field variation recorded at USGS observatories to obtain estimated time series of historical geoelectric fields. Integrating these geoelectric fields across power transmission lines results in time series of geomagnetically induced voltages on each power line. These voltages are analyzed statistically to construct hazard maps of the maximum voltages that could be realized in transmission lines across the US for an extreme, once in one hundred-year, geomagnetic storm. In combination with grounding resistance data and network topology, these voltage estimates can be utilized by power companies to estimate extreme geomagnetically-induced currents within their networks. These voltage estimates can provide information on which power lines and substations are most vulnerable to geomagnetic storms and can guide power companies in assessments of where to install additional protections within their grid.

The combination of strontium (87Sr/86Sr) and uranium (234U/238U) isotopes is an especially useful tool to track and quantify mixtures of water sources in arid wetlands, where chemistry and lighter isotopic tracers are strongly influenced by evaporation and transpiration. Those isotopes were used to understand modern water supply to the Pahranagat National Wildlife Refuge (PNWR) in southern Nevada (Paces & Wurster, 2014, J. Hydrol. 517). We investigate the possibility that this isotopic combination might also track paleohydrological changes in such settings. Here, we present Sr and U isotope data for authigenic carbonates in a sediment core spanning 5800 14C cal years from Lower Pahranagat Lake (LPAH), a shallow, alkaline lake within the PNWR. Modern surface waters in the PNWR are comprised of mixtures of discharge from two high volume springs from the regional carbonate aquifer in the northern valley, and smaller amounts of local discharge from the shallow volcanic aquifer in the southern valley. Modern surface water from LPAH has U isotopic values similar to the most recently formed LPAH carbonates; however, Sr isotopic values in LPAH surface water are somewhat lower than values in those same carbonates. Combined Sr and U isotope values in Holocene LPAH carbonates fall within the range defined by the three primary spring sources and reflect varying mixtures of those sources supplying LPAH from mid-Holocene to modern time. Values in the oldest samples (~5800 14C cal yr BP) have distinct 87Sr/86Sr and 234U/238U values that nearly match the local spring end member, suggesting that LPAH water at that time was dominated by proximal volcanic aquifer sources. By ~5300–5200 14C cal yr BP, LPAH water was comprised of a nearly equal mixture of the three spring sources, marking a dramatic shift in hydrologic conditions that allowed contributions of surface flow from distal carbonate springs to the north. Values in samples from ~1,000–3,000 14C cal yr BP indicate decreased contributions from distal carbonate springs during two drought intervals; however, by 730 14C cal yr BP, surface flow from carbonate springs had resumed. Our results indicate that combined Sr and U isotopes preserve evidence of past changes in water sources in arid wetlands useful for interpreting evolving hydrologic conditions.

Zonal flows in rotating systems have been previously shown to be suppressed by the imposition of a background magnetic field aligned with the direction of rotation. Understanding the physics behind the suppression may be important in systems found in astrophysical fluid dynamics, such as stellar interiors. However, the mechanism of suppression has not yet been explained. In the idealized setting of a magnetized beta plane, we provide a theoretical explanation that shows how magnetic fluctuations directly counteract the growth of weak zonal flows. Two distinct calculations yield consistent conclusions. The first, which is simpler and more physically transparent, extends the Kelvin-Orr shearing wave to include magnetic fields and shows that weak, long-wavelength shear flow organizes magnetic fluctuations to absorb energy from the mean flow. The second calculation, based on the quasilinear, statistical CE2 framework, is valid for arbitrary wavelength zonal flow and predicts a self-consistent growth rate of the zonal flow. We find that a background magnetic field suppresses zonal flow if the bare Alfvén frequency is comparable to or larger than the bare Rossby frequency. However, suppression can occur for even smaller magnetic field if the resistivity is sufficiently small enough to allow sizable magnetic fluctuations. Our calculations reproduce the η/B0^2 = const. scaling that describes the boundary of zonation, as found in previous work, and we explicitly link this scaling to the amplitude of magnetic fluctuations. These results could provide a plausible explanation for why the zonal jets in Jupiter go as deep as Juno has discovered but not any deeper.

Abstract: In Sichuan basin, the Longmaxi-Wufeng formation is rich of the shale gas, and the study of control factors of pores structure and connectivity are extremely significant for exploration of shale gas. Organic matter and inorganic matrix attribute to shale reservoir characters strongly. Though this work has been carried out by others before, it still has some arguments for this topic. In this project, we collected rock samples from oilfields, we did SEM, CO2 adsorption CH4 adsorption N2 adsorption, high pressure mercury injection, microscope observation, NMR, XRD, TOC test some related experiments. Analytical results show that TOC contents of the samples range from 0.96% to 6.12%, and the Wufeng-Longmaxi formation shale organic matter content is high. The samples porosity and permeability range from 0.49%-9.55% and 0.0006mD-0.73mD respectively, and the I/S of clay mineral amount ranges from 10%-60%. A plenty of graptolites were observed in the formation under microscope, and reflectivity of graptolites and solid bitumen can reveal that organic maturity is high in the formation. The higher the TOC percentage of sample, the higher the graptolite reflectance is. There are hydraulic fractures under microscope and SEM observation, and some of hydraulic fractures formed by stress, and some of them formed by hydrocarbon generation. Inorganic pores including bedding fracture, structural pores, intercrystalline pores, intergranular pores and dissolved pores and organic pore including corrosion pores, gas pores, bitumen spherulite pores can be observed. We observed quartz, calcite, feldspar under SEM to identify their cementation, filling, different periods’ growth, and aeolotropism. Based on these analysis data, authors inferred the shale diagenetic evolution is at the diagenetic B period, the mineral bitumen mass and solid bitumen macerals organic matters are contributed to the shale gas mainly, and by calculating the plane porosity by MATLAB, the fractal dimension is about 2.57. Pores diameters of these samples range from 1-55nm, and the maximum is 4nm. Most pores are around medium pores. In summary, shale gas migrates obeying non-Darcy flow along pores/fracture several nanometers, and gas migration affected by organic matter mainly in Wufeng-Longmaxi formation in Southeast Sichuan basin, China.

Empirical Mode Decomposition (EMD) is used to examine the relationship between precipitation and surface temperature from six regions. Three regions are defined by physiography: world, ocean, and land. The other three regions are defined by averaged precipitation: dry, normal and wet. Monthly averaged daily precipitation rate from the Global Precipitation Climatology Project are compared with average monthly surface air temperature anomalies from the Goddard Institute for Space Studies using EMD. The EMD process produces component time series referred to as intrinsic mode functions (IMFs). Theses IMFs are ordered by frequency from high to low. Eight IMFs were produced for each the time series. The first three IMFs corresponded to seasonal, semi-annual and annual variations, respectively. IMF 4 to 6 corresponded to a biennial, pentennial and decadal climate signals, respectively. IMF 7 was related to the broad 20-30 year period, with the trend being revealed in IMF 8. The time series spanned the period from January 1980 to December 2015 at monthly intervals. Temperature and precipitation time series from six sampling regions were analyzed for evidence of correlation. Results from the analysis reveal the following: (1) The EMD process reveals both linear and non-linear trends. The trends are not entirely consistent between regions though they are highly correlated. (2) Apparent wave-to-wave interactions between high and low frequency components appear to be observed in the IMF 1 and 2. These distortions appear to correspond to the troughs and peaks in the decadal cycle captured in IMF 6 and may related to the solar cycle. (3) The correlation between precipitation and temperature increases with increasing IMF number.

A new family of analytical magnetohydrostatic equilibria is presented. The methodology used to calculated these equilibria is based on work by Low (e.g. Low 1991) in the form given by Neukirch and Rastätter (1999). The equilibria have the capability of combining a non-force-free layer with a domain of linear force-free magnetic field, including a transition between these two domains. Potential applications to modelling solar magnetic fields will be discussed. Low, B.C., ApJ 370, 427 Neukirch, T. & Rastätter, L., A&A 348, 1000

The MASER (Measuring, Analysing and Simulating Radio Emissions) project provides a comprehensive infrastructure dedicated to low frequency radio emissions (typically < 50 to 100 MHz). The four main radio sources observed in this frequency are the Earth, the Sun, Jupiter and Saturn. They are observed either from ground (down to 10 MHz) or from space (down to a few kHz). Ground observatories are more sensitive than space observatories and capture high resolution data streams (up to a few TB per day for modern instruments). Conversely, space-borne instruments can observe below the ionospheric cut-off (10 MHz) and can be placed closer to the studied object. Several tools have been developed in the last decade for sharing space physcis data. Data visualization tools developed by the CDPP (http://cdpp.eu, Centre de Données de la Physique des Plasmas, in Toulouse, France) and the University of Iowa (Autoplot, http://autoplot.org) are available to display and analyse space physics time series and spectrograms. A planetary radio emission simulation software is developed in LESIA (ExPRES: Exoplanetary and Planetary Radio Emission Simulator). The VESPA (Virtual European Solar and Planetary Access) provides a search interface that allows to discover data of interest for scientific users, and is based on IVOA standards (astronomical International Virtual Observatory Alliance). The University of Iowa also develops Das2server that allows to distribute data with adjustable temporal resolution. MASER is making use of all these tools and standards to distribute datasets from space and ground radio instruments available from the Observatoire de Paris, the Station de Radioastronomie de Nançay and the CDPP deep archive. These datasets include Cassini/RPWS, STEREO/Waves, WIND/Waves, Ulysses/URAP, ISEE3/SBH, Voyager/PRA, Nançay Decameter Array (Routine, NewRoutine, JunoN), RadioJove archive, swedish Viking mission, Interball/POLRAD… MASER also includes a Python software library for reading raw data. This work is supported by CDPP, CNES, PADC and Europlanet-2020-RI. The Europlanet 2020 Research Infrastructure project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 654208.

We report the results of our investigations on ionospheric effects potentially caused by the 15 February 2013 Chelyabinsk meteor explosion. We used the observation data from a number of digisonde stations located in Europe and Russia to detect the traveling ionospheric disturbances (TIDs) likely to have been caused by the meteor explosion. We found that certain characteristic signatures of the TIDs can be identified in individual ionogram records, mostly in the form of Y‐forking/splitting of the ionogram traces. Based on the arrival times of the disturbances, we have inferred the overall propagation speed of the TIDs from Chelyabinsk to be 171 ± 14 m/s. In addition to the natural fulfillment of scientific endeavors, this work also highlights the importance of maintaining the mastery of ionosondes as ionospheric diagnostic instruments (in terms of operation, data analysis, representation, and interpretation) for many generations of space researchers to come.

The Younger Dryas (YD) in Bay of Islands, Newfoundland is characterized by major changes in both pollen and dinocyst assemblages, which translate into large drops in air and sea surface temperatures and sea surface salinity. Changes in vegetation are similar with those observed in Newfoundland lakes. Reconstructed air temperature shows a 7ºC drop at the time of the YD, an amplitude comparable to changes recorded in northeastern North America. Changes observed in dinoflagellate cyst assemblages are also like those at nearby sites from the Gulf of St. Lawrence and Laurentian Channel. Sea surface dropped by 7ºC, salinity by 4 psu and sea ice cover duration increased by 7 months. The YD period in eastern North America was traditionally defined by changes in sediments and pollen records. Our records show that the cooling in sea surface and air temperatures started 250 and 110 years before the start of the YD as defined by radiocarbon dates and pollen zones. By having exceptionally high resolution records for both terrestrial and marine conditions in the same core, we can show that there is a 140 years delay between ocean and atmosphere. We speculate that the major lithological and vegetation changes observed in lakes around the time of the YD period in eastern North America might actually represent the full impacts of the cooling, since the cooling trend actually started before the start of the YD period.

Formal data citation is a growing practice increasingly required by scientific journals. Roughly a decade ago, the Federation of Earth Science Information Partners (ESIP) began developing formal guidelines for data citation including acknowledgement of authors and archives and careful use of persistent identifiers (PIDs). Many Earth science data centers now provide formal citation text and PIDs for their data sets, typically a Digital Object Identifier (DOI). A central purpose of data citation (amongst many) is to aid scientific reproducibility through direct, unambiguous reference to the precise data used in a particular study, i.e., to aid provenance tracking. How has that worked in practice? ESIP is now in the process of revising and updating their guidelines and seeks to ensure that data citation meets its stated purpose. This presentation explores whether and how formal citation and the the use of PIDs for data sets has improved the tracking of data provenance. For example, is there is some commonality in the nature and granularity of objects that are assigned PIDs? We review how the guidelines are being revised to further enhance the transparency and reusability of data.

Reliable, accurate, and affordable linear motion systems for agricultural applications are currently not easily accessible due to their elevated cost. Most systems available to the public have price tags in the thousands and their dimensions cannot be easily customized. Current systems have a max length of about ten meters and for a typical greenhouse application, the length may not be sufficient. The price of the system increases with an increase in length and with a base price in the thousands it becomes almost impractical to buy a system for such application. The HyperRail is a modular linear motion system with a repeatability of 2mm and current top speed of 100mm/s. An advantage this system has is its ability to increase or decrease the length of system with minimum effort and nominal increase in price. The HyperRail can be mounted on a set of tripods or directly on the structure of a building such as a greenhouse. The base price for a three-meter system, on tripods, is US$240 and an additional US$45 for each additional one-and-a-half meter. The HyperRail was designed for the use of hyperspectral imaging but can be adapted for other sensor systems. We report on a nine-meter study over pine seedlings infected with a virus. A push-broom hyperspectral camera (Headwall Nano) was mounted on the carriage of the system imaging the seedlings. The rail is currently being adapted to an environmental sensor suite that will monitor CO2, luminosity, humidity, temperature, and the concentration of dust. The HyperRail also includes bidirectional-wireless communication between the drive and the carriage; this means that the sensor suite can operate autonomously and communicate to the HyperRail drive to move to a specific location and take measurements. This system includes a graphical user interface for users who are unfamiliar with programming but could also be used through a command line interface for individuals that want to work the code and see the effects of the changes immediately. This system was developed at Oregon State University’s OPEnS Lab, here is a link to the project page for more detailed information. URL for project page: http://www.open-sensing.org/hyper-rail/

The Piacenzian (Late Pliocene) represents a natural laboratory within which frequency and magnitude of environmental changes during a period of past global warmth can be analyzed, climate models can be tested, and results can be placed in a context to better prepare for future change. We focus on the North Atlantic and both generate new and incorporate existing faunal assemblage and alkenone data from DSDP Sites 552 and 606, ODP Sites 662, 981, 982, and 999, and IODP Sites 1307, 1308 and 1313. In addition, cores and outcrop material containing Piacenzian sediments from the Atlantic Coastal Plain of Virginia, USA, are included. These data allow us to characterize regional changes in temperature, salinity, upwelling, productivity, and diversity, associated with climate transitions, and make nuanced reconstructions of mid-Piacenzian conditions within a high-resolution temporal framework between ~3.40 and ~3.15 Ma, inclusive of Marine Isotope Stages M2 through KM5. Our results indicate the slope of the North Atlantic sea surface temperature gradient is a robust feature of the Late Pliocene and we confirm a temperature anomaly (Pliocene minus pre-industrial) little to no change at the equator to >6° C near 60° North, during Marine Isotope Stage (MIS) KM5c (3.21 Ma – 3.20 Ma). The M2-M1 transition is accompanied by warming, rapid sea-level rise, reorganization of planktonic foraminifer assemblages, decreases in productivity and increase in zooplankton diversity. Abundance changes in stenohaline taxa document repeated freshening of the Caribbean by low latitude Pacific water associated with brief re-openings of the Central American Seaway. The Gulf Stream – North Atlantic Drift migrated to a more northerly position during both the M2-M1 and KM6-KM5 transitions. Preliminary comparison of these results to climate simulations produced by the second phase of the Pliocene Model Inter-comparison Project (PlioMIP2) indicates model output and reconstructed environmental data are in closer agreement than previous data-model comparisons. We attribute this to a combination of improved Pliocene Research, Interpretation and Synoptic Mapping (PRISM4) boundary conditions, and the refined stratigraphic window used for verification, compared to PRISM3.