Creating a Time series NDVI map

Question

I would like to create a NDVI Time Series map? How do I do that? The one I probably would like to have similar like this http://proceedings.esri.com/library/userconf/proc02/pap0593/p05931.jpg

I created many NDVI for each or several months from the year of 2002 to 2013?

1) Do I need to add monthly NDVI to create an output monthly NDVI file and then the next step is to take new output monthly NDVI to produce a new output one year file?

(or)

2) Add all the NVDI from one year to generate a new output for one year file ?

The NDVI were produced from an ASTER.

Answer 1

When remote sensing vegetation, the time of year is very important. In most climates, vegetation has significantly more biomass (i.e., leaves etc.) during the summer, which means that it is easier for the sensor to discern the health of vegetation at that time of year. Two NDVI images of the same location from different times of the year may look different because they were taken at different points of the plants' growing cycles. For this reason, and because of differences in illumination, it is advisable to use images from the same time of year when developing a multi-year time series.

To answer your question, the image above seems to do this. All of the images are taken during the summer over a number of years. When developing your time series, you should pick a time of year that you have good data for (ideally near the peak of plant growth) and then create single NDVI images of that date for each year from 2002 to 2013. When developing your final product, you can put the images side by side like the image above. You do not create an NDVI for each month and then create a composite of them. For the aforementioned reasons, the images from each month of the year would look very different and a composite wouldn't be effective to achieve what you want.

The values outputted by NDVI indicate the relative health and abundance of the vegetation by comparing reflectance values in the red and near infrared bands. The maps above seems to be indicating that the health and abundance of vegetation in Mongolia increased between 1989 and 2001. The individual maps do not indicate change in plant growth over time; only collectively do they show how plant growth has changed over time.

Answer 2

Reproducing the map example you provided is primarily a cartographic effort and requires very little analysis if you have already calculated NDVI. I would use the following workflow to produce the map similar to the one you provided a link to.

1. Collect the NDVI data to use in your analysis. In the example, they use "Summer" 1989 to 2001. In your case, you would want to select a similar peak growth month (e.g. July) for the years 2002 - 2013.
2. Reclassify the NDVI values in the layer properties: right-click layer > Properties... > Classified or use Reclassify (Spatial Analyst). Reclassify July NDVI data to include breakpoints of interest to you. The example map uses 0.1, 0.2, 0.3, 0.4, 0.5, and 0.65 as thresholds.
3. Arrange your map to include 12 dataframes, one for each NDVI image. There is a short tutorial on How to create multiple synced data frames in ArcMap.
4. Finish by adding the usual cartographic features (e.g. legend, north arrow, credits etc).

I think it would be very helpful to also include a continuous raster surface showing the trend in NDVI values over time. There is a good tutorial on how to calculate a time series slope on GIS.SE. You would accomplish these calculation using the ArcGIS Raster Calculator (Spatial Analyst).

Answer 3

As you are trying to compare these images, the key is to do everything to make the images to be as comparable as they can be, so the actual changes can be seen.

Here are my steps:

1. Acquire Satellite images that are taken during the same time of the growing season, preferably peak biomass (see Christophers answer). Note, this might be some other time too, for example if you want to compare the onset of growing season you would acquire spring time images. This depends about the problem you are trying to answer with these maps.

2. Atmospheric conditions probably differ between images, so it is preferred to perform athmospheric correction. Please, see this paper by Song et al. 2001: Classification and Change Detection Using Landsat TM Data: When and How to Correct Atmospheric Effects?

3. Calculate NDVIs from atmo-corrected satellite images.

4. Classify NDVI values into same ranges. Breaks can be for example 0, 0.1, 0.2, 0.3, and so on. Make sure that breaks are the same in every image (as they are in your example .jpg). Classification and breaks really depends from your data. You propably need a little bit of trial and error to get it just right.

5. Set data frame layouts in ArcMap layout window and add north arrow and scale bar. Make a legend (one is enough when same classification is used).

Answer 4

If you need a service that will help you analyze, compare, track an area of your interest over a certain period of time then I would highly recommend to use LandViewer. This tool has a vast database of satellite images that are updated regularly.

All of this free satellite data can be found on LandViewer:

• Landsat 4 - archive 1982-1993
• Landsat 5 - archive 1984-2013
• Landsat 7 - archive since 1999
• MODIS - archive since 2012
• Landsat 8 - archive since 2013
• Sentinel-1 - archive since 2014
• Sentinel-2 - archive since 2015

All you need to do is choose a location, choose a date and apply NDVI index to the image. From there you may download the image in numerous formats and create a time-lapse.

Answer 5

From a remote sensing point of view, what you usually want in a time series is a consistent profile in time. Due to the presence of clouds, atmospheric degradation and BRDF effect (different illumination due to sun-object-sensor geometry), using single images will often result in noisy profiles. RS time series are therefore most of the time based on composite images : a combination of several images to reduce the noise and fill the gaps due to the clouds. For instance, maximum value compositing is used for MODIS and Mean Compositing is used for SPOT VGT. Those are simple method compared with BRDF correction methods, but Mean Compositing is very robust. You can have a look here for a weekly NDVI time series from SPOT VGT in Europe (using Mean Compositing).

From the GIS point of view, it depends on what you want to analysis/display. Yearly composite (all images from one single year) are useful for trend anlaysis. On the other hand, multi-year monthly composite (all images for the same month) are useful to assess the seasonality of the different land covers. Finally, one composite per month and per year gives you a full time series where you can observe changes in the seasonal patterns, but will probably be very noisy and with plenty of gaps if you are using ASTER data, so you will need temporal smoothing with plenty of issues. In your case, the example you show is more like a trend analysis (or multi-annual variability), so you should group summer images every year. If you have a large number of images each summer, I would try to make composite instead of picking just one image, because summer could start or end at different dates, so adding more images will bring more stability and remove the subjectivity of your choice for "the best". If you only have a few images, just picking the best one could be a solution (otherwise you will have some artefacts)

From the data processing point of view, it may help to create monthly composites then combine them into yearly composites (reduce the size in memory), but his is not necessary. However, if you do so with the mean compositing, you need to account for the number of valid observations each month and make a weighted average; otherwise your results will diverge.