Has the monsoon’s arrival date over my region shifted earlier or later?

What you can answer

• Onset day-of-year for each year over your AOI from an explicit, reproducible rainfall rule
• Multi-year trend in the onset date (Theil–Sen slope) and whether it’s significant (Mann–Kendall)
• Sensitivity of onset to the threshold and consecutive-day definition
• Circulation confirmation — whether monsoon westerlies / moisture flux (ERA5) accompany the rainfall jump
• Year-to-year variability and the spread of onset dates around the mean

What you can NOT answer with these alone

• The official IMD onset date — IMD uses station networks, winds and OLR criteria; satellite-only onset is an estimate, not a declaration.
• Long pre-2000 trends — IMERG begins in 2000; for multi-decadal records use gauge products (IMD gridded) or reanalysis, each with their own caveats.
• Withdrawal / break-monsoon timing without a separate rule.
• Causal attribution (why onset shifted) — requires climate-model and large-scale-driver analysis beyond these products.

Code template

import earthaccess
import xarray as xr
import numpy as np

earthaccess.login(strategy="netrc")

# AOI: Kerala (monsoon onset gateway)
aoi = (74, 8, 78, 13)
years = range(2001, 2025)

# --- Onset rule (IMD-style family — SET AND VERIFY for your region) ---
# Onset = first day on/after an earliest date where area-mean rainfall
# stays above a threshold for N consecutive days. Thresholds vary by
# definition; tune and report a couple of choices for robustness.
EARLIEST = "05-10"      # earliest permitted onset (mm-dd)
RAIN_MM  = 5.0          # daily area-mean rainfall threshold (mm/day) — verify
CONSEC   = 5            # consecutive wet days required — verify

def onset_doy(da_daily):
    """da_daily: 1-D xarray of area-mean daily rainfall for one year."""
    wet = da_daily >= RAIN_MM
    roll = wet.rolling(time=CONSEC).sum()  # count wet days in window
    hit = roll.where(roll == CONSEC, drop=True)
    return int(hit.time.dt.dayofyear[0]) if hit.size else np.nan

onsets = []
for yr in years:
    g = earthaccess.search_data(short_name="GPM_3IMERGDF",
                                bounding_box=aoi,
                                temporal=(f"{yr}-{EARLIEST}", f"{yr}-07-15"))
    # open -> area-weighted daily mean of 'precipitation' -> onset_doy()
    # optionally gate on ERA5 850 hPa westerlies / moisture flux here
    onsets.append(onset_doy(...))  # fill with computed series

# Trend: Theil–Sen slope + Mann–Kendall significance
import scipy.stats as ss
yrs = np.array(list(years)); od = np.array(onsets, float)
mask = ~np.isnan(od)
slope = ss.theilslopes(od[mask], yrs[mask])[0]   # days per year
tau, p = ss.kendalltau(yrs[mask], od[mask])      # MK-style monotonic test
print(f"trend = {slope*10:.2f} days/decade, p = {p:.3f}")

Expected output

• Series: onset day-of-year per year (2001–present), with the mean and ±1σ band
• Trend line: Theil–Sen slope in days/decade, annotated with Mann–Kendall p-value
• Sensitivity panel: onset dates under 2–3 threshold/consecutive-day definitions
• Optional: ERA5 850 hPa wind / moisture-flux composite around detected onset to confirm circulation
• Comparison markers: IMD official onset dates overlaid where available

Caveats

• Threshold sensitivity — onset can swing by many days with small changes to the rainfall threshold or consecutive-day count; always report the rule and test alternatives.
• Satellite rainfall bias — IMERG underestimates orographic rain over the Western Ghats and is uncertain for extremes; onset timing inherits this.
• Short record — ~25 years of IMERG limits statistical power; a slope can look like a shift yet have p > 0.05. Report the confidence interval.
• Bogus onset — strong pre-monsoon showers can trip a rainfall-only rule; gating on established westerlies/moisture (ERA5) reduces false onsets.
• IMERG latency / version — Final has ~3.5-month latency; use V07 and reprocess older cached granules.

Cross-DAAC composition

GES DISC (GPM IMERG, MERRA-2) + Copernicus/C3S (ERA5) — single DAAC if you stay on IMERG + MERRA-2; add a second auth only if pulling ERA5 from Copernicus.

Sources

• GPM IMERG: https://gpm.nasa.gov/data/imerg
• IMERG V07 user guide: https://gpm.nasa.gov/sites/default/files/2023-07/IMERG_V07_ReleaseNotes_230713-signed.pdf
• ERA5 (C3S): https://cds.climate.copernicus.eu/datasets/reanalysis-era5-single-levels
• IMD monsoon onset (Kerala) background: https://mausam.imd.gov.in/