Reading path: Underwater images and evidence →
Ask what produced the result
A marine paper may present observations, model outputs, or both. An observation is a record made from measurement or direct documentation, such as a sensor reading or an image-based count. A model result comes from calculations using a representation of a system and stated inputs or assumptions. Both can be informative, but they answer different kinds of questions.
Start with the methods section and figure captions. Look for terms such as “surveyed,” “measured,” or “sampled” alongside terms such as “simulated,” “predicted,” or “parameterized.” These clues are not a substitute for understanding the method, but they tell you where to look. A map of colored values might show measured stations, a modeled surface between them, or a scenario generated entirely by a simulation.
Follow data into the figure
For each important figure, identify the source of its plotted values. A time series could show instrument records, a model hindcast, or observations compared with model estimates. The caption may explain symbols or line styles, while the methods describe how inputs were processed. If those details are absent from the image alone, consult the main text before labeling the graphic an observation.
Consider a simple example: a study measures water conditions at several stations, then uses a model to estimate conditions between them. The station points are observations; the interpolated or simulated field is a derived product. A statement about the measured stations should not be worded as though every location was directly sampled. The method bridges the gap, but it does not erase it.
Understand assumptions and fit
For a model, ask which assumptions and inputs shape the output. Researchers may simplify processes, choose boundary conditions, or use parameter values from other work. These choices are part of the result’s context. A modeled scenario can show what follows if its assumptions hold; it does not by itself demonstrate that the scenario occurred in the ocean.
When observations are used to compare with model outputs, check how that comparison was made. Were the data collected at matching times and places? Does the paper report a measure of fit or describe mismatches? Agreement can support the model’s usefulness for a purpose, but it does not prove that all underlying mechanisms are correct. Disagreement may also reveal limits in measurements, inputs, or model structure.
Describe the evidence accurately
In your notes, label each key result as observed, modeled, or combined. Then state the setting and any assumptions that matter. For example: “Measured station values were used to evaluate a model’s estimate of the broader coastal pattern.” That phrasing tells readers which part comes from direct data and which part comes from computation.
Distinguishing evidence types helps you judge what a paper can establish. Observations document conditions sampled under a method; models explore relationships or scenarios that may be difficult to measure everywhere. When authors combine them, track the handoff from one to the other. That careful description keeps a useful model prediction from being mistaken for a direct field record.
Have a correction or a source to suggest? Send a note →