How are productivity maps generated?

Revision as of 09:14, 30 April 2014 by Mt (Talk | contribs)

Jump to: navigation, search

Remote Sensing basics

Please review the following articles, to get familiar with the basics on Remote Sensing for Agriculture. This is recommended to understand how Productivity Maps are generated

Methodology - combining NDVI from past seasons

Productivity maps are generated based on the combination of Vegetation Index maps, through geo-statistical analysis. It's is an automated and objective, that allows to compare the field production and variability in different times, in a single map.

Remote sensing imagery is selected in key dates, when crops are vigorous, and Vegetation Index (NDVI) maps are generated on those dates. These layers are then normalized, and integrated in the productivity map.  It's also possible to include information from previous yield monitor maps if available.

The following sample images and corresponding NDVI illustrate this process:

MZ landsat 8mar04.jpg
Landsat Image - Jan. 25, 2006
MZ NDVI 8mar04.jpg
NDVI - Jan. 25, 2006
MZ landsat 6jan05.jpg
Landsat Image - June 1, 2005
MZ NDVI 6jan05.jpg
NDVI - June 1, 2005
MZ landsat 25jan06.jpg

Landsat Image - Aug. 3, 2004

MZ NDVI 25jan06.jpg

NDVI - Aug. 3, 2004

These images are integrated by analyzing each pixel, defining homogeneous production areas, and obtaining the productivity map.
MZ management zone map.jpg

When is the right time to select imagery?

NDVI values are closely related to phenological vegetation patterns. The phenological pattern describes the functioning of vegetation and allows understanding
its geographic variations, while the NDVI determines vegetative activity.

the following figures show the relationship between NDVI evolution and crop phenology, in different research experiences

NDVI Phen1.jpg NDVI Phen2.jpg
Effect of nitrogen levels on NDVI (cotton) <ref>Spectral response of nitrogen fertilization in cotton (Gossypium species) Ansari, Mahey, et. al.</ref>

A second aspect to consider is that in the maturity stage, and depending on the type of crop, there is a close correlation between NDVI values and actual yields. Images to produce Productivity Maps are selected during that stage, to show yield variability.

Thus, it's very important to evaluate the Crop history of of the fields, when processing and interpreting NDVI maps, Productivity maps, or other layers derived from remote sensing processing. The crop planted in each field, with planting and harvesting date is key to know the right windows of time to analyze imagery.

It's also critical to know if and when there was any extraordinary event, such as hail, drought, floods, etc. Those years should be discarded, in order to avoid introducing a bias on the field's variability.

This is detailed in: Why is crop history important to analyze NDVI and Variability?

How often should Productivity Maps be updated?

Productivity maps can be generated anytime, but once started, this process can foster continuous improvement. By the end of the first season, and subsequent seasons, the results can be measured through yield monitor maps, or yield estimation through remote sensing. These additional layers can be used to improve the original productivity maps, and adjust the zones in the field; these updates can be done yearly or every other year to improve the knowdlege of the variability of the fields.


<references />