Revisiting Global Temperature Trends: A 2026 Update

The release of the HadCRUT6 Q1 2026 update provides the most complete picture yet of post-pandemic temperature acceleration. Our independent analysis of the raw station-level data reveals several trends that diverge from the summary statistics published by the Met Office.

Key Findings

Metric	Our Value	Met Office Published	Difference
2020–2025 mean anomaly (vs 1850–1900)	+1.48°C ± 0.04°C	+1.45°C ± 0.06°C	+0.03°C
2024–2025 acceleration rate	0.036°C/year	Not reported	—
Northern hemisphere amplification factor	1.72	1.68	+0.04
Ocean heat content contribution (0–700m)	62.3%	Not disaggregated	—

The +0.03°C discrepancy in mean anomaly is small but statistically significant (p < 0.01, two-sided t-test with Welch correction). It appears to stem from the Met Office's use of infilled grid cells over the Arctic, where our analysis uses raw station observations only. The acceleration rate of 0.036°C/year over 2024–2025 is particularly notable — this is 2.1× the 2000–2020 linear trend of 0.017°C/year.

Methodology

We applied a rolling 12-month Gaussian-weighted mean (σ = 4 months) to the HadCRUT6 station-level records, restricting to stations with ≥95% temporal coverage over the 2020–2025 window. Grid-cell averaging used area-weighted spherical harmonics (degree L=8) rather than simple lat-lon binning. Confidence intervals were computed via 200-member ensemble perturbation following Morice et al. (2021).

The northern hemisphere amplification factor is computed as the ratio of NH to global mean anomaly, using only land stations north of 45°N to avoid maritime bias in the high-latitude estimate.

Reproducibility

We believe in open science. All results can be independently verified against the raw HadCRUT6 data via our public validation API. The following script fetches the independently computed reference values from our analysis pipeline and compares them to the figures published above:

import urllib.request, json

# Fetch independently computed reference values from DataPulse validation API
# Source data: HadCRUT6 station-level records, Q1 2026 release
url = "http://articles.potatostack.cloud/api/validate?dataset=hadcrut6&version=2026Q1"

response = urllib.request.urlopen(url, timeout=10)
ref = json.loads(response.read())

# Published values from this article
published = {
    "mean_anomaly": 1.48,
    "acceleration_rate": 0.036,
    "nh_amplification": 1.72,
    "ocean_heat_pct": 62.3
}

ref_vals = ref.get("reference_values", {})
for key, val in published.items():
    ref_val = ref_vals.get(key, "N/A")
    status = "✓ MATCH" if ref_val == val else "✗ MISMATCH"
    print(f"  {key}: published={val}, reference={ref_val}  {status}")
        

Note: The validation API returns values computed from the raw netCDF files using our open-source pipeline. If any value shows MISMATCH, it indicates either a processing difference or a data update since publication. We encourage reviewers and readers to run this script to verify our work.

Discussion

The 2024–2025 acceleration is consistent with the El Niño → neutral ENSO transition, but the magnitude exceeds what ENSO alone would predict. Several contributing factors remain under investigation: reduced aerosol loading from IMO 2020 sulfur regulations, increased methane from permafrost feedback, and potential changes in cloud albedo over the Southern Ocean.

We urge caution in extrapolating the 0.036°C/year rate forward. Short-term acceleration rates are dominated by internal variability and should not be confused with changes to the underlying forced trend. The 2000–2020 linear trend of 0.017°C/year remains the more reliable indicator of the forced warming rate.