Datasetpublic

juliensimon/mars-chemcam-compositions

Mars ChemCam LIBS Oxide Compositions Part of the Planetary Science Datasets collection on Hugging Face. Major oxide compositions of Mars surface rock and soil targets analyzed by the Chemistry and Camera (ChemCam) Laser-Induced Breakdown Spectroscopy (LIBS) instrument aboard the Curiosity rover. Currently 30,458 individual point analyses across 4,184 named targets, spanning sols 0 to 4612. Dataset description ChemCam fires a focused laser pulse at rock and soil… See the full description on the dataset page: https://huggingface.co/datasets/juliensimon/mars-chemcam-compositions.

sourceHugging Facecc-by-4.0updated 6mo agoView on Hugging Face

0likes64downloads

Dataset Card

Mars ChemCam LIBS Oxide Compositions

Part of the [Planetary Science Datasets](https://huggingface.co/collections/juliensimon/planetary-science-datasets-68228b04b65e1f3b9e57a76b) collection on Hugging Face.

Major oxide compositions of Mars surface rock and soil targets analyzed by the Chemistry and Camera (ChemCam) Laser-Induced Breakdown Spectroscopy (LIBS) instrument aboard the Curiosity rover. Currently 30,458 individual point analyses across 4,184 named targets, spanning sols 0 to 4612.

Dataset description

ChemCam fires a focused laser pulse at rock and soil targets up to ~7 meters away, creating a plasma whose emission spectrum reveals elemental composition. The Major Oxide Compositions (MOC) data product provides predicted weight percentages for nine major oxides (SiO2, TiO2, Al2O3, FeOT, MgO, CaO, Na2O, K2O, MnO) derived from the LIBS spectra using a combined PLS+ICA multivariate model.

Each row represents a single LIBS analysis point. Multiple points are typically measured per target to characterize compositional variability. Uncertainty estimates (RMSEP and shot-to-shot standard deviation) are provided for each oxide.

Schema

Column	Type	Description
`file`	string	Source spectrum filename
`target`	string	Named target on the Martian surface
`sio2`	float64	Silicon dioxide (wt%)
`sio2_rmsep`	float64	SiO2 RMSEP uncertainty from calibration model
`sio2_shots_stdev`	float64	SiO2 shot-to-shot standard deviation
`tio2`	float64	Titanium dioxide (wt%)
`tio2_rmsep`	float64	TiO2 RMSEP uncertainty from calibration model
`tio2_shots_stdev`	float64	TiO2 shot-to-shot standard deviation
`al2o3`	float64	Aluminum oxide (wt%)
`al2o3_rmsep`	float64	Al2O3 RMSEP uncertainty from calibration model
`al2o3_shots_stdev`	float64	Al2O3 shot-to-shot standard deviation
`feot`	float64	Total iron as FeO (wt%)
`feot_rmsep`	float64	FeOT RMSEP uncertainty from calibration model
`feot_shots_stdev`	float64	FeOT shot-to-shot standard deviation
`mgo`	float64	Magnesium oxide (wt%)
`mgo_rmsep`	float64	MgO RMSEP uncertainty from calibration model
`mgo_shots_stdev`	float64	MgO shot-to-shot standard deviation
`cao`	float64	Calcium oxide (wt%)
`cao_rmsep`	float64	CaO RMSEP uncertainty from calibration model
`cao_shots_stdev`	float64	CaO shot-to-shot standard deviation
`na2o`	float64	Sodium oxide (wt%)
`na2o_rmsep`	float64	Na2O RMSEP uncertainty from calibration model
`na2o_shots_stdev`	float64	Na2O shot-to-shot standard deviation
`k2o`	float64	Potassium oxide (wt%)
`k2o_rmsep`	float64	K2O RMSEP uncertainty from calibration model
`k2o_shots_stdev`	float64	K2O shot-to-shot standard deviation
`mno`	float64	Manganese oxide (wt%)
`mno_rmsep`	float64	MnO RMSEP uncertainty from calibration model
`mno_shots_stdev`	float64	MnO shot-to-shot standard deviation
`sum_of_oxides`	float64	Sum of all oxide compositions (wt%)
`distance_m`	float64	Distance from rover to target (meters)
`laser_power`	string	Laser power settings
`spectrum_total`	string	Total spectrum intensity
`sol_range_min`	int64	Start of sol range for this data file
`sol_range_max`	int64	End of sol range for this data file

Quick stats

—30,458 LIBS point analyses across 4,184 targets
—Sols 0 to 4612 of Curiosity's traverse
—Mean SiO2: 47.6 wt% | Mean FeOT: 18.0 wt%

Usage

python

from datasets import load_dataset

ds = load_dataset("juliensimon/mars-chemcam-compositions", split="train")
df = ds.to_pandas()

# Average oxide composition per target
target_avg = df.groupby("target")[
    ["sio2", "tio2", "al2o3", "feot", "mgo", "cao", "na2o", "k2o"]
].mean()

# High-silica targets (possible felsic rocks)
felsic = target_avg[target_avg["sio2"] > 60].sort_values("sio2", ascending=False)

# Iron-rich targets
iron_rich = df[df["feot"] > 25].sort_values("feot", ascending=False)

# Composition variability within a single target
target_std = df.groupby("target")[["sio2", "feot", "mgo"]].std()

Data source

PDS Geosciences Node — MSL ChemCam LIBS RDR, Washington University in St. Louis. Major Oxide Compositions (MOC) derived using the combined PLS+ICA multivariate model (sPDL Tool v2.5).

Related datasets

—mars-craters — Robbins Mars crater catalog
—neo-close-approaches — Near-Earth object approaches
—small-body-database — JPL small body orbital parameters

Pipeline

Source code: juliensimon/space-datasets

Citation

bibtex

@dataset{mars_chemcam_compositions,
  author = {Simon, Julien},
  title = {Mars ChemCam LIBS Oxide Compositions},
  year = {2026},
  publisher = {Hugging Face},
  url = {https://huggingface.co/datasets/juliensimon/mars-chemcam-compositions},
  note = {Based on MSL ChemCam LIBS MOC data from the PDS Geosciences Node}
}

License

CC-BY-4.0