1

I have downloaded five datasets and would like to quantify the rate of urban sprawl in a watershed from 2001 to 2011 using NLCD. Here is the data I have: (1) NLCD 2001 to 2011 Land Cover from to Change Index, (2) NLCD 2011 Land Cover (3) NLCD 2011 Percent Developed Imperviousness, (4) NLCD 2001 Land Cover (2011 Edition), and (5) NLCD 2001 Percent Developed Imperviousness (2011 Edition).

Once I clip the data to my watershed boundary, how do I calculate the rate of urban sprawl?

I know that I won’t need all of the above data, but figured this is essentially everything I need for my study.

All of the data was downloaded from https://www.mrlc.gov/index.php

0

First, you need to take some decisions. Do you want the rate year by year or just from the beginning to the end?

Second, classify the land covers by only two categories according to what you consider urban and non urban. Urban can be reclassified to 1 and non urban to 0. (Search for "reclassify raster" if you don't know how to do it). Use the layers that you think are more relevant for your index. Once you have your rasters with only two categories, urban and non urban you can start doing some stat.

So, now imagine you have the raster from 2001, you just need to count the number of urban cells and non urban cells using raster statistics, (In Arcmap is under the properties menu). If you're happy with descriptive statistics, you just need to pass these data into Excel or any other statistics package to calculate the rate.

5
  • Thank you! I will calculate total change in percent impervious cover from 2001-2011 and use only 2001 and 2011 Land Cover datasets. These datasets have many blank classes and few colored classes (impervious etc.) a total of about 95, should I delete the blank classes? Once I have two maps (2001 and 2011) that each have rasters with only two categories, can I combine them and create a map that displays the change from 2001-2011? Will I be able to calculate area in sq km using raster statistics?
    – Mike
    Jul 15 '17 at 3:00
  • If you use the raster calculator to add both maps you can do what you are asking me. Change first layer values to 1 and 2 and second one to 3 and 4. Multiply 1x3=3 it was non urban and stayed the same. 1x4=4 it was non urban and changed to urban. 2x3=6 it was urban and changed to non urban. 2x4=8 it was urban and stayed the same. Then, you can change 3,4,6,8 to other value that make more sense to you. Don't forget to write down what they mean or you'll go crazy later on. Update metadata. The final raster after this multiplication will be the map you want and you will be able to check the stats
    – Sergio C.
    Jul 15 '17 at 3:15
  • Please mark the answer as correct if you think it helped you.
    – Sergio C.
    Jul 15 '17 at 3:16
  • Awesome, I marked it as correct. One more quick question - the NLCD is projected in albers_conical_equal_area and my watershed shp is wgs_1984_utm_zone_16n, do I need to reproject anything? The map looks correct without reprojecting. Thanks again.
    – Mike
    Jul 15 '17 at 13:34
  • Arcmap and other softwares adjust the projection so it looks fine but it can still lead to errors. Reproject the wgs 1984 to Albers conical so they are in the same projwction
    – Sergio C.
    Jul 15 '17 at 17:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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