CO2 PPM - Monthly Mauna Loa
- Rows
- 823
- Columns
- 6
- Data range
- 1958-03-01 to 2026-09-01
- Auto-refresh
- Re-checks its source every 14 days
- Last refreshed with new data
- 2026-11-12
Monthly mean atmospheric CO2 concentrations at Mauna Loa, Hawaii. Long-term time series from 1958 to present, science/environmental data.
AI analysis
CO2 PPM Monthly Mauna Loa — Dataset Analysis
1. Dataset Overview & Value for Correlation Studies This dataset captures monthly mean atmospheric CO2 concentrations recorded at Mauna Loa Observatory in Hawaii, spanning March 1958 through April 2026 — a remarkable 68-year continuous time series representing one of the most scientifically significant environmental records in existence. Sourced from [datahub.io/core/co2-ppm](https://datahub.io/core/co2-ppm/r/-/data/co2-mm-mlo.csv), a curated open data publisher known for maintaining core reference datasets, the record extends to 2026, suggesting active maintenance and near-current freshness. This longevity and regularity make it exceptionally well-suited for long-term trend analysis and cross-domain correlation studies linking atmospheric chemistry to economic, ecological, and climate variables.
2. Data Quality Observations Data quality is notably strong across the board. All 818 rows are fully populated with zero null cells across all 6 columns, eliminating any need for imputation or missing-value handling. The Date column carries 818 distinct values matching the row count exactly, confirming no duplicate monthly records — each row represents a unique month. The Average column has 794 distinct values versus 818 rows, and Interpolated has 786 distinct values, indicating a small number of repeated measurement values, which is expected given the gradual pace of atmospheric change and measurement precision limits. The most notable quality flag is the Number of Days column, which reports 205 outliers — a disproportionately high count (~25% of rows) — and contains negative values as low as -9.99, suggesting this column may encode a data-quality or completeness indicator rather than a literal day count, and should be treated with caution or excluded from naive numerical analysis.
3. Key Column Distributions The Average CO2 column (in PPM) tells the core story: a mean of 361.02 PPM and median of 356.26 PPM, with a range of 312.42 to 431.12 PPM. The positive skew of 0.379 reflects the accelerating upward trend in recent decades pulling the distribution rightward. The interquartile range (Q1=331.12, Q3=387.29) spans ~56 PPM, capturing the bulk of mid-century to recent readings. The Interpolated column mirrors Average almost identically (mean=361.02, σ=33.13, skew=0.38), confirming it serves as a gap-filled surrogate with minimal divergence. The Trend column is the most anomalous structurally — typed as Integer with values ranging only -1 to 31 and just 25 distinct values, a negative skew of -0.961, and a median of 25 — strongly suggesting it represents a derived categorical or ordinal indicator (possibly a data-quality or smoothing flag) rather than a continuous trend value, and warrants metadata clarification before use in regression.
4. Recommended Join Key Columns The Date column is the clear primary join key, offering 818 fully distinct, null-free monthly timestamps from 1958-03-01 to 2026-04-01 in ISO date format — ideal for time-series alignment. The Decimal Date column (min=1958.2, max=2026.29, mean=1992.25) provides a continuous numeric alternative useful for mathematical modeling or joining datasets that encode time as fractional years. For any cross-dataset join, monthly or annual granularity alignment should be validated, as some partner datasets may only offer annual observations, requiring aggregation of this monthly series.
5. Suggested Paired Datasets for Correlation Discovery Several dataset categories would pair powerfully with this record. Global surface temperature anomalies (e.g., NASA GISS or NOAA datasets) are the most natural pairing, enabling direct CO2-warming correlation analysis across the same time window. Global fossil fuel consumption and industrial emissions data (e.g., Our World in Data, IEA) would allow source-attribution analysis, testing whether emission spikes lead CO2 readings by predictable lag periods. Oceanic pH / ocean acidification datasets would extend the chemistry story downstream. Deforestation and land-use change records could reveal whether terrestrial carbon sink disruptions correlate with PPM acceleration. Finally, economic indicators such as global GDP, energy intensity, or recession periods (e.g., World Bank time series) could illuminate how economic contractions — visible as dips in the 1970s oil crisis or 2008–2009 recession — leave detectable signals in the CO2 trend.
Columns
- Date (date)
- Decimal Date (decimal)
- Average (decimal)
- Interpolated (decimal)
- Trend (integer)
- Number of Days (decimal)