Deciduous fruit crop estimate methodology - CROP ESTIMATE by sdsdfqw21



                          December 2003

              Dr. Daan Louw and Mariette Fourie
     Optimal Agricultural Business Systems cc [97/144487/23]
               258 Main Street, PO Box 163, Paarl, 7622
             Tel: +27 (21) 8702953, Fax: +27 (21) 8702915
                         Cellular: 0828573458

The motto of SA Table Grapes, "Grapes from South Africa…always fresh, always
sweet", might just as well apply to any of our SA export fruits. Given the current state of
information, which is critical for short medium and long-term planning, logistic planning
for the future has become a nightmare for almost all the role-players in the fruit supply
chain. If the SA fruit industry cannot rectify this problem soon the primary objective of
always fresh, always sweet will not be reached. The following abstract from a recent
letter by the National Ports Authority of South Africa indicates their concerns:

"The congestion in the Cape Town Container Terminal (CTCT) is a point of concern for
all our customers. Delays caused as a result thereof has widespread consequences for all
port users. The CTCT is presently experiencing unprecedented growth in volumes. This
combined with present infrastructure layout, has led to a situation where the CTCT finds
itself at the limit of its design capacity".

This is only a fraction of the problems experienced due to a lack of reliable information.
Analysing these problems can in almost all cases be traced back to the lack of a relatively
accurate crop estimate. An accurate crop estimate is probably the most important
component of any strategy to improve the efficiency of the logistical process. While
realising that it is impossible to make a 100% accurate crop estimate, the current state of
the estimate is unacceptable.

The advantages of a relative good crop estimate - before the season and with frequent
updates - must be obvious:

 Input providers such as packaging, chemical and fertilizer companies will normally
   do their planning to carry a minimum volume of stocks. They plan their stocks based
   on crop estimates. In the event of a bad crop estimate excessive stocks or a shortage
   will increase the price of these inputs. Who bears these costs? Ultimately the primary
 The transport industry will be able to do better logistic planning to ensure that there
   are no delays in transporting our products from the packing/cooling facilities to the
   export harbours, mainly Cape Town. If the crop estimate is way out, they will either
   have excess capacity or a shortage of capacity in the main production regions. Who
   bears this cost? The producer will bare most of it.
 With inaccurate crop estimates the container industry will either have excess
   containers or there will be a shortage. In both cases this will increase costs. Who
   bears these costs? Ultimately the primary producer.
 Inadequate or excess shipping capacity can create tremendous problems for the
   industry. Not only does this contribute to increased costs but it also creates
   bottlenecks for the harbour authorities by causing delays. In most cases the producer
   will ultimately bear these costs.
 Exporters will be able to do better market co-ordination. Who bears the cost of bad
   co-ordination? Ultimately the producer.

The consequences mentioned above are only part of the problem. These problems
normally occur during a season. What about medium to long-term crop estimates? The
objective with medium to long-term crop estimates is mainly focused on the planning of
infrastructure to enable the supply chain to handle larger volumes. At present the authors
are not aware of any scientific long-term crop estimates. In most cases, whether it is
government or private sector, institutions will attempt to delay large capital expenditures
for as long as possible. Also, there is a time delay from approval of these projects until
construction starts. In most cases it can take months or years to create additional
capacities such as roads and harbour terminals. In the absence of relatively accurate
medium to long-term crop estimates the industry will soon find itself in a crisis to ship
our fruit timely without quality problems and within international protocols to our export

The objective with this document is to give an international overview of crop estimate
methodology and to apply what can be learned to propose a methodology

An internet survey on crop estimate methodology for fruit did not reveal much. Although
there are many references on the internet for the actual estimates, information on the
methodology is surprisingly limited. The following is a summary of what was found in
selected countries:

Australia (Citrus)
The approach followed is to develop a uniform methodology for a crop
forecasting/estimation system and planting statistics database for the Australian citrus
crop and to encourage its implementation throughout all citrus producing regions within

A national citrus estimate database and crop forecasting methodology is already in place
in the three major citrus growing regions (Riverland, Sunraysia, Riverina) in conjunction
with the Australian Citrus Growers Federation (ASGF).

The crop forecasting methodology consists of employing three field officers for crop
forecasting that are all using the same methodology and data collection forms for density
counts and fruit size measurements.

Their Citrus Database is a relational database management system using the Microsoft
Office suite of programs operating in the Microsoft Windows environment. It is a system
of crop forecasting based on establishing a variation to previous season's crop utilizing
count and size measurements and planting and other statistics. The citrus database
    grower survey records (confidential);
    planting statistics;
    fruit density count and measurement statistics;
    packer and processor throughput information to give harvest update and market
      distribution analysis.

Experience in Australia and other countries have shown that mandarins are often more
difficult to forecast accurately due to their cropping variability.

The crop supply and demand estimates prepared by the United States Department of
Agriculture (USDA) are crucial to both policy makers in government and people
involved in making decisions about marketing and investing. In today’s information age,
statistical methods provide the benchmark against which all other data sources are
compared. The agencies involved should have a solid record of objectivity. The Secretary
of Agriculture’s office is briefed about the results only after the final results have been
completed and prepared for distribution to the public. The security of the data before
release is fiercely defended with extraordinary efforts to ensure there is no premature
disclosure of any of the information.

The World Agricultural Outlook Board (WAOB) coordinates an interagency process that
prepares monthly forecasts of supply and demand for major crops, both for the United
States and the world, and follows a balance-sheet approach to account for supplies and
utilization. The major components of the supply and demand balance sheet are beginning
stocks, production, domestic use, trade, and end-of-season carryout stocks. Whereas
forecasts of US crop production and estimates of US stocks on hand are independently
prepared by National Agricultural Statistics Service 9NASS), US and foreign supply and
demand forecasts are developed jointly by several USDA agencies. Unfortunately only
major staple foods are covered by their system with no co-ordination for horticultural
crops. The authors are of the opinion that this is a major shortcoming for world trade in
fruit and vegetables.

Crop production forecasts have two components, namely area to be harvested and
expected yield per ha. The first forecasts are based on conditions as of the survey
reference date and projected assuming normal conditions for the remainder of the season.
For example, the assumption of "normal conditions" is that temperatures and
precipitation will be at historic averages for the remainder of the season. For consecutive
estimates, the crop maturity and conditions at the reference date are evaluated against the
time remaining. Long-range weather projections are not used as an indicator for final

The review process followed to develop consecutive crop estimate updates involves
evaluating the relative ranges of the forecast errors of the grower yields and the
objective measurement of yields and the degree to which they overlap. If there is a
significant change in conditions between the survey period and the report date such as a
killing freeze, serious heat wave, beneficial rains, etc., the primary goal is to provide the
most accurate production forecast possible given the available survey data. The official
estimate may represent a departure from the survey averages, but will still reflect the
current crop conditions within the ranges provided by the data. When forecasting crop
yields, NASS, does not attempt to predict future weather conditions. Long-range
weather forecasts are not used in any forecast models. To the extent that conditions

depart from normal, the forecasts also will fluctuate. Procedures used to prepare area
estimates and yield forecasts are discussed in the following sections.

The methodology proposed in this document is not new, but rather an effort to bring
consistency into the crop estimate methodology. It should be clear that consistency in
methodology should be the key to get confidence in any estimation methodology. It has
been shown in Australia, the USA and also in South Africa that it is possible to estimate
the crop within less than a 5% deviation from the actual figures by using historical actual
production, estimates of new plantings, fruit set information from technical experts and
estimates of crop damage should they occur.

The discussion that follows is generic for all fruit. The concepts are the same for all fruit
kinds but will definitely need adaptation to cater for the specific character of different
fruit kinds. The approach followed was to develop a uniform methodology for a crop
forecasting/estimation system for all fruit. The methodology was applied to table grapes
and stone fruit for the 2003/2004 season. This is discussed in the following section
which deals with the application of the methodology.

The proposed       methodology/procedure     should    include   at least the     following

    It must preferably be part of the activities of a marketing forum which has
      representatives from producers, exporters, local market and technical experts who
      can give an opinion on production circumstances in the major production regions
      of the country. In the absence of a forum, the producers’ organisation must take
      responsibility for the crop estimate.
    The service provider that conducts the crop estimate should be independent and
    There should be an ongoing awareness program that emphasizes the importance of
      a proper crop estimate.
    First estimate (theoretical estimate): During the first estimate it is not possible to
      make an accurate assessment of fruit sizes and/or sometimes even fruit set since it
      is too early. In the case of fruit kinds where the majority (80% or more) are being
      exported it is possible to use the average of the previous three season’s actual
      intake inspection information as base estimate. The purpose of this calculation is to
      reduce the impact of natural deviations (good/bad seasons). The frequency
      (weekly/monthly/total) will depend on the fruit kind. For example, it is adequate to
      estimate apple and pear for the total production and marketing channel only since
      it is possible to stretch the marketing season over several months. However, in the
      case of table grapes and stone fruit it is necessary to do the estimate on a per
      cultivar and per week basis and also in some cases per fruit size. Tree/vine census
      information can be used to estimate expected growth/decrease in production for the
      different production regions. The initial estimate can be done without any
      participation from stakeholders.

    After the establishment of a first, theoretical crop estimate the results should be
      discussed with technical representatives to provide inputs (e.g. on cold units in the
      winter preceding the season and frost damage) that can shed more light on
      reductions/increases in the estimate and/or early/late season impacts. After such a
      discussion the first estimate could be adapted if necessary and it is then imported to
      publish a first official crop estimate.
    Second crop estimate: In the second estimate the process is easier since an
      estimate is already available. However, at this stage (normally a month after the
      first estimate) it is possible to add more detail to the estimate in terms of fruit sizes
      where applicable and there will be more certainty about the fruit set in different
      regions. A pre-determined pro-forma for crop estimates (which is established at the
      marketing forum or through consultation with the industry) is then sent out to
      exporters/pack houses (depending on the fruit kind). The target should be to cover
      at least 80% of total production. It is important for participants to stick to deadlines
      to send back the information. Since, in most cases, it is not possible to cover 100%
      of the production volume it is necessary to extrapolate (blow up to 100%) the data
      to the total volume. The method of extrapolation will depend on the circumstances
      but will in general consist of a procedure to spread the extrapolated volume in such
      a way as to provide for differences in regions, sub-sectors within a fruit sector and
      differences in cultivars.
    Consecutive estimates: The total number of estimates within each fruit sector will
      depend on cultivar spread. In most cases it will be necessary to conduct a third or
      even fourth estimate depending on the length of the production season.

The methodology described above was applied to stone fruit and table grapes for            the
2003/2004 season. At the time of writing it was too early in the season to evaluate        the
results. However, the results of the first few weeks are provided. To illustrate           the
theoretical crop estimate the table grape industry will be discussed and to discuss        the
second crop estimate stone fruit is used as case study.

Table grapes: theoretical crop estimate
Since the database in its existing format is only complete for two years, it was necessary
to calculate the average production over only two seasons instead of the preferably three

                Theoretical table grape crop estimate for the 2003/2004 season


                                                                                      2001/2002 Season
  2,500,000                                                                           2002/2003 Season





               44 45 46 47 48 49 50 51 52 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23

Also, it was assumed that growth in the major production regions will not be the same.
The following was assumed based on vine census information adapted for natural
disasters during the previous season.

 Region                                % increase in production
 Berg River                                     +0%
 Hex Valley                                     +2%
 Limpopo Province                               +5%
 Olifants River                                 +0%
 Orange River                                   +12%

Stone fruit
Stone fruit consists of apricots, peaches, nectarines and plums. The actual crop estimates
received from exporters represented 74% of the total stone fruit exports based on last
season's exports and therefore a 26% extrapolation was done on these volumes. The
extrapolation process was as follows:

1)      The average exports per stone fruit group (Peaches, Nectarines, Apricots and
        Plums) over the last 3 years were calculated.
2)      The average total stone fruit exports for the last 3 years were calculated.
3)      The average exports per stone fruit group is divided by the average total stone fruit
        exports for the last three years to calculate each stone fruit group's contribution to
        total stone fruit exports.
4)      Then the stone fruit group’s contribution to the total stone fruit exports was
        multiplied with the total amount of stone fruit cartons extrapolated to get the
        extrapolation per stone fruit group.
5)      Thereafter the crop estimate (as supplied by the exporters) weekly volumes were
        taken, weighted and multiplied with the extrapolation total per stone fruit group to
        get the extrapolated volume per week.
6)      Extrapolation was calculated by inflating the crop estimate received from the
        exporters with 26% (Shortfall % based on last year's exports).

After the first estimate was presented and discussed at the SFJMF meeting it was decided
to shift the whole estimate forward with one week due the late nature of the season. Only
the results for apricots are presented in this report. The following graph shows the
estimation process for apricots.

                                                           Apricots: crop estimate 2003/2004







                    40   41      42     43    44     45     46    47    48    49     50       51   52        1     2     3     4    5      6     7      8      9     10     11     12       13

                              2000/ 2001 Act ual Cartons         2001/2002 A ctual Cart ons            2002/ 2003 Act ual Cartons       2nd Est St andard cart on equivalent s 2003/ 2004

It is clear that a three-year average on export volumes per week gives a good base to start
from. During the second estimate the marketing forum indicated that the season was
approximately 10 days later than the first estimate and it was decided to shift the estimate
forward with one week. Although the peak week for the estimate and the intakes
coincide, the intake volumes only reached the estimated weekly volumes in week 49.
Thus the form of the graph is correct, but the weekly intake volumes were over-

                                                      Inspections Passed for export

  4,75 equivalent cartons






                                      44         45     46      47          48       49      50    51         52      1

                             GRANDIR APRICOT          BULIDA APRICOTS            EXE. APRICOTS          CHARISMA APRICOTS
                             PALSTEYN APRICOTS        Estimate

Although it will be necessary to review the methodology from time to time, the
researchers are positive that this methodology is accurate enough for decision making


To top