Sentinel 5P Aerosol Index outside range

Hi, I am trying to download Sentinel5P data but I noticed the values for layers AER_AI_354_388 and AER_AI_340_380 are outside the normal range specified here. It is reported that the minimum value for these layers is -1 but I often encounter values lower than this threshold. I am downloading the data using Python SDK in this way

request = SentinelHubRequest(

                evalscript=evalscript,

                input_data=[

                    SentinelHubRequest.input_data(

                        data_collection=DataCollection.SENTINEL5P,

                        time_interval=(ts, ts),

                        mosaicking_order=MosaickingOrder.MOST_RECENT,

                        upsampling="BICUBIC"

                    )

                ],

                responses=[SentinelHubRequest.output_response("default", MimeType.TIFF)],

                bbox=city_bbox,

                size=city_size,

                config=config,

            )

            img = request.get_data()[0]

where evalscript is the following

function setup() {

    return {

        input: [{

            bands: ["AER_AI_354_388", "dataMask"],

        }],

        output: {

            bands: 2,

            sampleType: "FLOAT32"

        }

    };

}



function evaluatePixel(sample) {

    return [

        sample.AER_AI_354_388,

        sample.dataMask

    ]

}

the same for AER_AI_340_380 . I very rarely noticed some negative values also for HCHO layer which uses the exact same script.

Am I doing something wrong ?

Thanks in advance for the help

Page 1 / 1

Hi @links ,

Your script looks fine to me. As mentioned in the documentation, the Typical Range indicates what values are common for a given band and unit, however outliers can be expected.

I randomly downloaded an original file from Copernicus and the histogram did indicate that there’re values lower than -1 (Fig 1).

Fig 1

Hi @chung.horng, thanks for your reply.

I understand that Typical Range indicates common values but negative values do not make so much sense to me since they should be concentration values in mol/m^2. I would then consider those values as wrong acquisition and discard them, what do you think?

In addition, I noticed that the tile downloaded with my script and the values reported in EO Browser are different. More in detail, considering the metropolitan area of Milan on 2018-04-30 for some regions my downloaded tile has values lower than -1 while EO Browser shows values around 0.5 in the same area.

For this reason, I am still forced to think there is something wrong with downloaded data, or the alternative could be that the EO browser does some post-processing before visualisation. What do you think about this?

Thanks again for your support

Hi @links ,

Could you provide your aoi and the timestamp which shows the discrepancy?

Hi @chung.horng,

my AOI is ( 9.0408867, 45.3867381, 9.2781103, 45.5358482 ).

Regarding AER_AI_340_380 and AER_AI_354_388 layers the timestamp in which I observed the discrepancy between my tile and EO Browser is 2018-04-30T10-27-52.

For what concerns HCHO layers I have to check the timestamp in which I obtained negative values but I can easily discard them as negative concentrations are for sure due to a noisy acquisition.

Thanks for your support

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.