2

I am trying to remote sensing land cover features (bare soil, water, vegetation, cultivated areas) on a 40 km² watershed, located in the Sahel (semi-arid west africa). I am heavily interested in their spatio-temporal trends and patterns. As such, I found Landsat images to be potentially good candidates. Free available Landsat products (C1 and C2 levels) have a resolution of 30 m, i.e the surface being represented in a single pixel is close to 900 m².

I would like to know if this spatial resolution is accurate enough for my task, given the size of my study area (40 km²) ?

Is there a general rule of thumb to follow, or even better, some references discussing the topic ?

How can I account for and estimate mapping errors subsequently ?

closed as too broad by PolyGeo Feb 21 '18 at 16:37

Please edit the question to limit it to a specific problem with enough detail to identify an adequate answer. Avoid asking multiple distinct questions at once. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.

  • 2
    Assuming a 6.3km x 6.3km study area, I think that works to a square of a little over 200x200 landsat pixels. I think you can certainly use this for spatiotemporal tends as each image will have the same footprint and a regular period....however, the classification result will look rather pixelated for sure and I am not sure that thisfootprint will offer you enough sample sites to define your training sites for classification. – Zipper1365 Feb 21 '18 at 14:00
6

The size of the area is not the primary issue for the selection of a satellite sensor resolution. You should rather think about the size of the objects that you are mapping and their contrast with the background (see this paper about sub-pixel detection). My "rule of thumb" is that you do not detect accurately an object that is less than four pixels if you don't have some extra knowledge about its background.

In most parts of the United States, you can make quite accurate land cover maps with Landsat-8 or the likes. In Sahel, however,crop fields are very small and heterogeneous, and the vegetation is very sparse, so it is difficult to capture them very well. Sentinel-2 data is therefore a better option if you only need recent years, but even Sentinel-2 data might not be good enough in your case.

As mentionned by @Zipper1365 , rendering could be another issue. For rendering, the rule of thumb is that pixels should not exceed 0.3 mm on the map. So the scale to display your data would be 1/100 000 (e.g. 5 by 8 cm for the printed map of 40 km²).

  • Thanks for the swift and learningful answer. The paper you mentionned sounded a bit "technical" for my understanding (provided I have not a strong background knowledge on the topic) but it adressed the core point. Thanks. – Roland Y. Feb 21 '18 at 14:21
  • I actually wanted to reach people from UCL for further discussion regarding my research work. In case @radouxju is interested and available, please let me know how we can get in touch. Thanks! – Roland Y. Feb 21 '18 at 14:35
  • @RolandY. you can send me an email – radouxju Feb 23 '18 at 10:23

Not the answer you're looking for? Browse other questions tagged or ask your own question.