SatHDSS
Network panel

Algae Dataset

Every data hub, dataset, sensor, variable, paper and limitation in one tree — and for each: where it comes from, how far back it goes, how often it captures, why it belongs in a cholera study, and why it does not work everywhere.

Highlight site All three Chakaria HDSS Matlab HDSS Dhaka City Choosing a site greys out everything that does not apply there.

2. Inland/coastal high-res

AL05 · Landsat 5 / 7 / 8 / 9 Collection 2 Level-2
USGS / NASA
A+ Gold FREE NO ACCOUNT

Source · duration · capture cadence

Data source
USGS / NASA
Access route
GEE · USGS M2M API · AWS Open Data S3
Endpoint / asset
LANDSAT/LT05|LE07|LC08|LC09/C02/T1_L2 | m2m.cr.usgs.gov/api
Duration
1984 (L5) / 1999 (L7) / 2013 (L8 → present
Capture cadence
16 days per satellite (~8 d with two in orbit)
Resolution
30 m
Delivers
Chlorophyll proxy (green/blue, needs local calibration), FAI, turbidity, MNDWI, water surface temperature
Command
10_algae_daily_3sites.py --sensors landsat
Why it is used for cholera. Well validated for land · water-quality retrieval requires LOCAL calibration
Why not — catalogue limitation. NO RED-EDGE BAND → true NDCI is impossible. This is the weakest link in your 2000-2015 record. Get Sarker et al. 2019 (Front. Mar. Sci. 6:598) for Bangladeshi coefficients. L7 SLC-off stripes after 2003-05-31.

Site applicability

SiteVerdictWhy
Chakaria HDSS YES ESSENTIAL for 2000-2015
Matlab HDSS YES ESSENTIAL for 2000-2015
Dhaka City YES ESSENTIAL for 2000-2015

Reliable · accessible · authentic

A+ Gold
Reliability tier
OPEN
Accessible
USGS
Authentic — operator

Typed limitations (3)

HIGH 16-day revisit misses short blooms entirely temporal

One Landsat satellite passes every 16 days. A bloom that initiates, peaks and collapses inside two weeks can fall entirely between two passes and leave no trace in the record.

What to do: Two satellites in orbit halve the effective interval to about 8 days. Combine L7+L5 for 2000-2012 and L8+L9 for 2021 onward. Report bloom FREQUENCY with an explicit detection-probability caveat.

HIGH Landsat 7 SLC-off stripes after May 2003 coverage

The scan-line corrector failed in 2003, leaving wedge-shaped data gaps that widen toward the scene edge and lose roughly 22 percent of each image.

What to do: Mask the gaps rather than gap-filling them, and record the usable water-pixel count per pass. The water_px column already does this.

HIGH No red-edge band, so no NDCI spectral

Landsat has no band near 705 nm. The green/blue ratio used instead is the open-ocean approach and is known to fail in turbid productive water - which is every water body in this study.

What to do: Treat the Landsat chlorophyll series as a RELATIVE index only, and never splice it to a Sentinel-2 NDCI series without the bias correction in 11_algae_daily_fill.py.