Earthdata Forum

Hi everyone,

I'm trying to access Level-3 chlorophyll-a data from NASA's PACE mission (PACE_OCI_L3M_CHL_NRT) using the earthaccess Python package. The response indicates no granules are found:

Details:
App: Python-based ingestion pipeline using earthaccess, xarray, pandas, and pymysql

Data Product: PACE_OCI_L3M_CHL_NRT

Dataset: Level-3 Chlorophyll-a concentration

Data Variable: chlor_a (automatically parsed from NetCDF)

Access Method: earthaccess.search_data() with these parameters:

short_name="PACE_OCI_L3M_CHL_NRT"

temporal=("2024-12-15", "2025-01-07")

bounding_box=(-61.5, -53.2, -57.5, -50.9) (Falkland Islands region)

cloud_cover=(0, 50)

Platform: Accessing via NASA Earthdata Cloud (earthaccess)

Operating System: Windows 11

Earthaccess Version: (latest pip install)



Would really appreciate guidance — is the PACE_OCI_L3M_CHL_NRT dataset not available yet for this region/date, or am I structuring the request wrong?

Current code:

# 📁 config.py
import pathlib

# Earthaccess will handle authentication
username =
password =

# This will prompt once and store credentials in ~/.netrc
import earthaccess
auth = earthaccess.login(persist=True)

# Region of interest: Chesapeake Bay area (guaranteed to have data)
bounding_box = (-61.5, -53.2, -57.5, -50.9)
start_date = "2024-12-15"
end_date = "2025-01-07"
iron_release_date = "2024-12-30"

# Specific OCI L3M chlorophyll product
product_list = ["PACE_OCI_L3M_CHL_NRT"]


# Output directory for downloads
download_dir = pathlib.Path("downloads")
download_dir.mkdir(exist_ok=True)

# Cloud cover range (optional filter)
cloud_cover = (0, 50)

# MySQL DB config (DigitalOcean)
db_config =


# 📁 ingest.py
import os
import xarray as xr
import pandas as pd
from config import product_list, bounding_box, cloud_cover, download_dir, iron_release_date , start_date, end_date
import earthaccess

def fetch_and_process():
all_metrics = [] # store results here

for product in product_list:
print(f"🔍 Searching for granules for: {product}")

# Search with Earthaccess
results = earthaccess.search_data(
short_name=product,
temporal=(start_date, end_date),
bounding_box=bounding_box,
cloud_cover=cloud_cover
)

if not results:
print(f"⚠️ No granules found for {product}")
continue

print(f"📥 Downloading {len(results)} granules...")
paths = earthaccess.download(results, download_dir) # download locally

for file_path in paths:
try:
ds = xr.open_dataset(file_path)
date = file_path.name.split(".")[1][:8] # extract date from filename
period = "before" if date < iron_release_date.replace("-", "") else "after"

for var in ds.data_vars:
data = ds[var]
all_metrics.append({
"product": product,
"variable": var,
"filename": file_path.name,
"date": pd.to_datetime(date),
"period": period,
"mean": float(data.mean(skipna=True).values),
"max": float(data.max(skipna=True).values)
})
except Exception as e:
print(f"⚠️ Failed to read {file_path.name}: {e}")

os.remove(file_path) # clean up

return pd.DataFrame(all_metrics)

if __name__ == "__main__":
df = fetch_and_process()
print(df.head()) # Optional: preview results




Thanks in advance!
— Will

The time period you are referencing should have refined CHL data, not NRT. Remove NRT from your datatype and retry your query.

Tommy

Hey there,

Unfortuntly i try this and still get 503 Server Error: Service Temporarily Unavailable.

server at obdaac-tea.earthdatacloud.nasa.gov is not currently accepting connections for the files and i try querying multiple times in the day and different days and i get nothing.

Hi Will, were you able to retrieve your data or are you still getting a 503 error?

do you think it makes sense to get on a call, if so do you have an time and date you are available as well as your email to send an invite :)

Hey, still getting error, can we jump on a call. let me know what date/ time suits you and what time zone youre from :)

A 503 error indicates a server-side error, which may be intermittent. I was able to run your code (after editing it to restore proper indentation - in the future you can wrap the text in the "code" element - see the angle bracket icon above -to preserve indentation). I did have to comment out the cloud_cover option in the earthaccess.search_data call, since the L3m collection doesn't report cloud percentage, but then it work as expected. (as an aside, I did notice there are other issues with your code after the earthaccess dowload that are not relevant to the access issue.)

If you are still getting 503s, send your IP address in an email to connection_problems@oceancolor.sci.gsfc.nasa.gov and our network folks can see if there is an issue with the connection (like a potential block).

Regards,
Sean