Distribution of Gymnodinium catenatum Graham cysts and its relation to harmful algae blooms in the northern Gulf of California

Germination of cysts serves as inoculum for the proliferation of some dinoflagellates, and cyst abundance in sediments represents crucial information to understand and possibly predict Harmful Algae Blooms (HABs). Cyst distribution is related to the physical characteristics of the sediments and the hydrodynamics (circulation) of a particular region. In the northern Gulf of California (nGC) several Gymnodinium catenatum HABs have been recorded. However, the presence of resting cysts and the effect of hydrodynamics on their distribution in the nGC have not been investigated. This study evaluated cyst abundance, distribution and their relation to local circulation in surface sediments during two periods that coincided with a non-bloom year condition (July 2016) and after a major HAB registered in the nGC that occurred in January 2017. Also, a numerical ocean model was implemented to characterize the transport and relocation of cysts and sediments in the region. Gymnodinium catenatum cysts were heterogeneously distributed with some areas of high accumulation (as high as 158 cyst g−1, and 27% of total cyst registered). Cysts seemed to be transported in an eastward direction after deposition and accumulated in an extensive area that probably is the seedbed responsible for the initiation of HABs in the region. The nGC is a retention area of cysts (and sediments) that permit the formation of seedbeds that could be important for G. catenatum HAB development. Our results provide key information to understand G. catenatum ecology and specifically, to understand the geographic and temporal appearance of HABs in the nGC.

Relationship between Three-Dimensional Radiation Stress and Vortex-Force Representations

In numerical ocean models, the effect of waves on currents is usually expressed by either vortex-force or radiation stress representations. In this paper, the differences and similarities between those two representations are investigated in detail in conditions of both conservative and nonconservative waves. In addition, comparisons between different sets of equations of mean motion that apply different representations of wave-induced forcing terms are included. The comparisons are useful for selecting a suitable numerical ocean model to simulate the mean current in conditions of waves combined with currents.

Parallel I∕O in Flexible Modelling System (FMS) and Modular Ocean Model 5 (MOM5)

We present an implementation of parallel I∕O in the Modular Ocean Model (MOM), a numerical ocean model used for climate forecasting, and determine its optimal performance over a range of tuning parameters. Our implementation uses the parallel API of the netCDF library, and we investigate the potential bottlenecks associated with the model configuration, netCDF implementation, the underpinning MPI-IO library/implementations and Lustre filesystem. We investigate the performance of a global 0.25∘ resolution model using 240 and 960 CPUs. The best performance is observed when we limit the number of contiguous I∕O domains on each compute node and assign one MPI rank to aggregate and to write the data from each node, while ensuring that all nodes participate in writing this data to our Lustre filesystem. These best-performance configurations are then applied to a higher 0.1∘ resolution global model using 720 and 1440 CPUs, where we observe even greater performance improvements. In all cases, the tuned parallel I∕O implementation achieves much faster write speeds relative to serial single-file I∕O, with write speeds up to 60 times faster at higher resolutions. Under the constraints outlined above, we observe that the performance scales as the number of compute nodes and I∕O aggregators are increased, ensuring the continued scalability of I∕O-intensive MOM5 model runs that will be used in our next-generation higher-resolution simulations.

The Vertical Structure of Open-Ocean Submesoscale Variability during a Full Seasonal Cycle

Submesoscale dynamics are typically intensified at boundaries and assumed to weaken below the mixed layer in the open ocean. Here, we assess both the seasonality and the vertical distribution of submesoscale motions in an open-ocean region of the northeast Atlantic. Second-order structure functions, or variance in properties separated by distance, are calculated from submesoscale-resolving ocean glider and mooring observations, as well as a 1/48° numerical ocean model. This dataset combines a temporal coverage that extends through a full seasonal cycle, a horizontal resolution that captures spatial scales as small as 1 km, and vertical sampling that provides near-continuous coverage over the upper 1000 m. While kinetic and potential energies undergo a seasonal cycle, being largest during the winter, structure function slopes, influenced by dynamical characteristics, do not exhibit a strong seasonality. Furthermore, structure function slopes show weak vertical variations; there is not a strong change in properties across the base of the mixed layer. Additionally, we compare the observations to output from a high-resolution numerical model. The model does not represent variability associated with superinertial motions and does not capture an observed reduction in submesoscale kinetic energy that occurs throughout the water column in spring. Overall, these results suggest that the transfer of mixed layer submesoscale variability down to depths below the traditionally defined mixed layer is important throughout the weakly stratified subpolar mode waters.

Parallel I/O in FMS and MOM5

We present an implementation of parallel I/O in the Modular Ocean Model (MOM), a numerical ocean model used for climate forecasting, and determine its optimal performance over a range of tuning parameters. Our implementation uses the parallel API of the netCDF library, and we investigate the potential bottlenecks associated with the model configuration, netCDF implementation, the underpinning MPI-IO library/implementations and Lustre filesystem. We investigate the performance of a global 0.25° resolution model using 240 and 960 CPUs. The best performance is observed when we limit the number of contiguous I/O domains on each compute node and assign one MPI rank to aggregate and write the data from each node, while ensuring that all nodes participate in writing this data to our Lustre filesystem. These best performance configurations are then applied to a higher 0.1° resolution global model using 720 and 1440 CPUs, where we observe even greater performance improvements. In all cases, the tuned parallel I/O implementation achieves much faster write speeds relative to serial single-file I/O, with write speeds up to 60 times faster at higher resolutions. Under the constraints outlined above, we observe that the performance scales as the number of compute nodes and I/O aggregators are increased, ensuring the continued scalability of I/O-intensive MOM5 model runs that will be used in our next generation higher resolution simulations.