Where can I find land cover data for Kerala from 2008 - 2012?

[Shijith Kunhitty]

So I'm doing some work on Kerala, and wanted to use a land cover dataset that covers any year from 2008 to 2012. 

Does anyone have any idea on where I could find such data? Preferably at 30m resolution?

I was thinking of using one of those global land cover datasets, but they don't work for one reason or the other.

I know there is Glance from Boston University that covers 2001-2019, but they've only released data for North America so far.

There are some datasets from academics in China -- FROM-GLC 30, Globeland30 -- but their websites are either difficult to navigate or don't work anymore.

There are other datasets from NASA and ESA, but at 500m and 300m their spatial resolution isn't high enough.

If anyone could give me any leads, I'd really appreciate it

Thanks,

Shijith

[Devdatta Tengshe]

Hi  Shijith,

Your requirement of a 30 m Land Cover from 2008 to 2012 is a difficult ask.

This is because there wasn't a good multispectral dataset of that vintage which was widely available. It is only with the advent of the Sentinel System, and advances in storage, that generating a Global, or even county wide LULC dataset became reasonable, at such a high resolution. ( A rule of thumb is that your source Multispectral data should be 1/3 spatial resolution of your required LULC)

Having said that, there are 2 options that you could look into:

1) ESRI has released a Global LULC of 2019-2020 vintage which is freely available here: (https://livingatlas.arcgis.com/landcover/)

2) NRSC/Bhuvan has a LULC from 2005 Data, which was made at a 1:50k resolution, which is approximately equivalent to 50m resolution. The bad news is that this is not directly downloadable, and you will have to write to them, and see if they are willing to share the data with you.

A third option is that if you have enough resources, you could download/buy data of your study area, and generate the LULC.

Regards,

Devdatta

--
Datameet is a community of Data Science enthusiasts in India. Know more about us by visiting http://datameet.org
---
You received this message because you are subscribed to the Google Groups "datameet" group.
To unsubscribe from this group and stop receiving emails from it, send an email to datameet+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/datameet/105f2d6b-de1b-4fad-9c44-c1c2a201c357n%40googlegroups.com.

[Sarath Guttikunda]

One more set of lulc datasets - 1992 to 2020

https://urbanemissions.info/india-air-quality/india-ncap-landcover-classification/

Portal has files cut for Indian city airsheds and the link to global files is at the top

With best wishes,

Sarath

--
Dr. Sarath Guttikunda
http://www.urbanemissions.info

To view this discussion on the web visit https://groups.google.com/d/msgid/datameet/CAOQOUMu_yXjMBpjYc-YTjrkh3uf8BqUWXiHGKUkx3%3DkqU0Hbxw%40mail.gmail.com.

[Pradeep]

A yet another freely available land cover dataset, for 2020, at 10m, from ESA, is WorldCover: viewable at https://viewer.esa-worldcover.org/worldcover/, more information at https://esa-worldcover.org/.

[Shijith Kunhitty]

Thanks Devdatta, Sarath and Pradeep for responding.

Managed to find land cover data for Kerala from Bhuvan's WMS service. Unfortunately, they only provide png/jpeg images, so had to write code to convert those images into the underlying data. Just sharing a link to the python code in case others want it.

(Also I work in the media and these calculations gives me a rough idea of landuse/landcover, so that's good enough for me. But if you're in academia, i think the actual data may be available for purchase, please do that instead.)

The code should give you LULC data at 1:50,000 scale for any area in India for 2005-06, 2011-12 or 2015-16 and save it as an netcdf file. 

In the netcdf file, the LULC class will be under the band 'class_new'. What each class code represents can be seen from this bhuvan request. And for visualising the data, use the 'red_new', 'green_new' and 'blue_new' bands. 

Have pasted a few lines of the code below. 

regards, shijith

# docstring for function

def download_lulc_data(state, bbox, period, save_path):
    '''
    This function gets 1:50,000 or 25m resolution land use/land cover imagery \n
    from bhuvan's wms service, estimates the underlying data, and \n
    saves that data as a netcdf file.

    Parameters:
    state -- location of area of interest. Use 2-letter codes like KL for Kerala
    bbox -- bounding box for area of interest: [min_lon, min_lat, max_lon, max_lat]
    period -- 3 options are '2005-06', '2011-12' and '2015-16'
    save_path -- where netcdf will be saved
    '''

# example of usage
state = 'KL'
bbox = (76.5368557, 9.5704093, 76.5883541, 9.628581)
period = '2011-12'
save_path = 'kl_2011_12_lulc.nc'
download_lulc_data(state, bbox, period, save_path)