Thursday, 12 May 2016

NDVI based Site Selection for Crop Cutting Experiments

Satellite Derived NDVI based Site Selection for Crop Cutting Experiments
Introduction:

Agricultural statistics has great importance for the planners for planning of most important economic policies for betterment of the people of the country.  Crop area and crop production forms the backbone of any agricultural statistics system.  In India, crop area figures are compiled on the basis of complete enumeration while the crop yield is estimated on the basis of sample survey approach.  The estimates of yields are obtained on the basis of scientifically designed Crop Cutting Experiments (CCE) conducted under a scheme of the Directorate of Economics and Statistics, Ministry of Agriculture (DESMOA entitled “General Crop Estimation Surveys” (GCES).

Yield:

The yield estimates of major crops are obtained through analysis of Crop Cutting Experiments (CCE) conducted under scientifically designed General Crop Estimation Surveys (GCES). At present over 95% of the production of food grains in India is estimated on the basis of yield rates obtained from the CCEs. The CCEs consist of identification and marking of experimental plots of a specified size and shape in a selected field on the principle of random sampling, harvesting and threshing the produce and recording of the harvested produce for determining the percentage recovery of dry grains or the marketable form of the produce.

CCE:

Initially 1,70,000 CCE were planned at all India level to generate district-wise estimates of yield for major crops with less than 5% of margin of errors. For a few crops like Paddy, Wheat etc, the size of the sample was sufficient enough to generate estimate at block/tehsil level. Gradually, the number of CCEs increased to accommodate more number of crops under GCES and recently for the purpose of crop insurance to generate yield estimate at Gram Panchayat level. At present at all India level approximately 8,88,000 CCEs are conducted.

Limitations of to conduct these many CCEs spends much time and effort in scrutinizing the data for gaps in coverage and inconsistencies, seeking clarifications from different organizations and first hand impressions from state officials and field visits, and auxiliary information on trends in input use (for yields). This exercise, and adjustments to estimates of area and yield reported by states based on them, is focused mainly on rice, wheat and a few other major crops. The Directorate needs to be commended for the care and seriousness with which they try to do as best as they can in presenting estimates.

·        But inevitably informed, but necessarily subjective, judgment plays an important role in this process. Most assessments of trends in agricultural production and policy discussions use the estimates as the best available and broadly consistent with other macro economic trends. But the margins of error remain uncertain. The need for improvements in the quality of the basic data and for more transparent and objective ways of making adjustments to imperfect data using cross checks with independent sources and consistency with the behavior of consumption, prices and wages is indisputable.

·         The deficiencies in the current system of both area and yield estimation and are not due to deficiencies in its design. The selection of sample villages for collecting data on land use and crop area, as well as the selection of sample villages and plots for crop cutting experiments are based on rigorous and statistically sound principles.

WITH THE GROWING TECHNOLOGY AND TECHNIQUES MOST OF THE ABOVE SAID ISSUES OR LIMITATIONS CAN BE SOLVED BY USING REMOTE SENSING AND GIS TECHNIQUES.


The identification of Plot numbers on the ground is an issue. Farmers also interfere in deciding the experimental plot. To overcome all these problems, the satellite-based index values (NDVI) come in handy.

The choice of representative NDVI range was made by identifying the concentration of adjacent 60% - 80% of the values from the median (30% or 40% on either side). The median is an ideal choice of averaging the yield in such cases as some large values might affect the mean. Mode that depends on frequency has been ruled out.


Low NDVI represents low yielding areas
High NDVI represents high yielding areas
Median NDVI represents medium yielding areas




Thursday, 5 May 2016

Remote sensing helps farmers and policy makers

Nice Article and it is true

Remote sensing has helped the national crop estimates committee (CEC) achieve a 0.13% error margin in 2015 from 9.77% in 2002 and will allow policy makers to assess drought impact down to municipal level in the Western Cape.
When the agricultural control boards such as the Maize Board and Wheat Board were abolished in 1997 there was a need to have an independent body to provide crop estimates to guide farmers, processors, futures dealers and policy makers. This led to the establishment of the Crop Estimates Committee (CEC) within the Department of Agriculture. To prevent conflict of interest no member of the CEC could trade in the commodities being forecast.
The initial crop estimates were collected by means of a telephonic survey of industry participants, be they farmers, processers or agricultural cooperatives. As the graph below shows, there was an inherent bias to under-estimate the national maize crop in part because low supply would push up prices.
1
To help reduce the bias, the producer independent crop estimate system (PICES) was developed which uses remote sensing using satellite images and aerial surveys using low-flying aircraft and helicopters to estimate acreage planted to the various crops. The remote sensing is so sophisticated that it can distinguish between the various plant types and asses how well they are growing. This helps with the yield estimates, which are also compiled by on-site surveys conducted by the state-run Agricultural Research Centres (ARC) Grain Crops Institute.
From 2009 to 2015, the CEC had three underestimates and four overestimates for the size of the crop so there was no prevalent bias. In addition, the error margin has moved from 9.77% in 2002 to only 0.13% in 2015, a feat better than most national crop estimates conducted in the world, where the error margin is “satisfactory” if it is less than 5%.
In particular, the CEC was justifiably proud of its record in the last three years when despite wide fluctuation in the maize crop from 11.69 million tonnes in 2013 to 14.25 million tonnes in 2014 to 9.94 million in 2015, the error margins were respectively -.13%, +0.4% and -0.75%.
“The fact that we had less than one percent error in the past three seasons shows how robust the system is as during that time the maize crop slipped by 3.6% in 2013 from 2012, then expanded by 21.9% to the largest crop since 1981 before plunging by 30.2% as the drought cut yields,” said Eugene du Preez, director of privately-held SiQ, which provides the committee with satellite and aerial data, which helps it determine the size of the area planted.
The expertise and equipment built up to service the CEC has allowed SiQ to expand its services so that it can help individual farmers and provincial departments of agriculture. Recently they have conducted surveys to see how much arable land in the former homelands of Ciskei and Transkei could be used for maize, while also conducting census for other provincial departments.
One such census was recently completed in the Western Cape and that showed that there 97 crop types cultivated and that there was a massive 12,000 facilities in terms of agricultural infrastructure such as abattoirs. The remote sensing could answer questions such as how many pear orchards are there and what is the amount of hectares devoted to pears. What is the value of crops per municipality and how does this fluctuate in good and poor weather conditions? How many farms are there devoted to game and how much to sheep or cattle? What agri-tourism facilities are there and what do they offer?
This kind of detail means that SiQ can say that the top five districts in terms of grain production potential in the Western Cape are the Swartland at R757m, Cape Agulhas at R585m, Thewaterskloof at R547m, Hessequa at R465m and Swellendam at R463m. This will help policy makers with assessing how much revenue is likely to decrease due to drought given that Western Cape dams are at critical levels with Voelvlei dam for instance at only 19% of capacity.

Courtesy: CAPE Business news

Popular Posts