Yield forecast

The yield prognosis based on vegetation indices derived from NOAA.AVHRR images has been carried out at the Center of Remote Sensing IGiK. The information about indices have been included into the yield modelling system. For every ten-day period of the vegetation season the forecast of yield reduction for each NUTS2 division has been assesed.

Now, the yields of cereales we forecast are below averages. The reason is the lack of rain and high temperature which lasts from the beginning of May. It couses the extensive agricultural drought in Poland. The maps showing the coverage and degree of drought on agricultural area for the following decades we present in the section Drought detection.

In Geoinformation Issues, 2011 Vol. 3 No 1(3), is published new article entitled:
Monitoring of agricultural drought in Poland using data derived from environmental satellite images.
It is available on the website of Digital Libraryof IGiK.


Course of yield reduction forecast

Our models predict crops yield with average error on the level of 5%. We analysed the reduction of yields in relation to the potential ones in the  region:
(Ymax-Yforec)/Yforec. On the right side forecasts of yield reduction  from 10th to 20th ten-day period  were presented. Three years were selected due to their diversity. Year 2012 was extremely good, 2010 was very wet with average yields, 2006 was very bad with poor frozen cereals period and later drought. Each decadal prognosis is compared to the results published by CSO.

At the Department of Remote Sensing the works aimed at yield forecasts were carried out within research projects (grant)

Grant No. 4 T12E 026 30
Grant No. 5 T12E 026 23
Geoland Sub-project II
PR-3






 

IGIK Forecast error CSO Result
21.8% 2006
6.9% 2010
15.0% 2012
decade :

Yield forecasting system

The Central Statistical Office (CSO) does assessment of potential yield of main crops every year. The method of visual crop condition evaluation in the preliminary and pre-final assessments of crop yield is being influenced to a considerable extent with subjective approach. In addition, the procedure of data collection in the field as well as processing them is labour and time consuming. Therefore, the Remote Sensing Department have undertaken the research to explore remote sensing methods of crop growth conditions and yield assessments. The objective of the work is to develop the yield forecasting system.

The satellite images (NOAA AVHRR ; SPOT VEGETATION) have been used as the source of information on crop growth conditions and yield forecasting for cereals. The indices - the Vegetation Condition Index (VCI), the Accumulated Vegetation Condition Index (AVCI) - both based on the NDVI - and the Temperature Condition Index (TCI) - based on surface temperature - have been computed for each ten-day period of the year for the entire agriculture areas in Poland.

These indices have been continually correlated with cereal yield and critical periods of crop development were found as significant for this relationship. The information about indices from particular ten-day periods have been included into the cereals and winter wheat yield modelling. The database of the indices and statistical models obtained were used for developing the Predictive Indices (PTVCI, VCIAVG, and VTCIAVG) applied for the forecast of yield for each ten-day period of the year. The method have been used operationally for the crop forecast.

The results of our research are illustrated on graphs presenting the yield forecasting process in years 1999 - 2004 for cereals and winter wheat and an example of comparison between very good for agriculture year 2004 and the year with drought 2006.

On the left: the yield forecasting system elaborated at the Department of Remote Sensing IGiK.

10-days raster maps and maps of mean values for polish voivodeships are delivered to Department of Agriculture of Central Statistic Office of Poland