Monday, June 15, 2009

Checking the quality of your leaf analysis data


Fertilizer costs (materials, application) represent more than 80% of variable costs in a plantation. To achieve the maximum economic yield, it is therefore important that the recommended application rates for each fertilizer in every planted field is as close as possible to the site's nutrient requirement.

As a guide for the preparation of fertilizer recommendations, most oil palm plantations rely on leaf analysis, in combination with field assessments and an analysis of historical trends in leaf nutrient status, nutrient inputs, and yield.

Considering the cost of mineral fertilizer inputs, however, agronomists should take the precaution of making crosschecks on the laboratory to confirm that the variabilities within a single batch or between different batches is indeed due to differences in palm nutrient status, and not due to errors in laboratory analysis. It is only possible to carry out crosschecks if dried and ground leaf samples have been submitted to the laboratory for analysis. We recommend that each estate of >100,000 ha purchase a leaf grinder so that this can be carried out.

First, using normal sampling procedures, collect sufficient Frond #17 leaflets from healthy palms to produce 10 kg of dried, ground leaf material. This sample then becomes the standard estate sample (SES) and should be thoroughly mixed and stored in an airtight glass container in a refrigerator. Ten samples from the SES should be submitted to a reputable laboratory to determine reference values for all future standard analyses.

Whenever leaf samples are submitted to the laboratory for analysis, include ten SES as a crosscheck. These SES samples should be indistinguishable, in terms of labelling, from the normal samples. When the results are available from the laboratory, enter the results from the SES samples in a spreadsheet and check for variability (Table 1).

Table 1. Mean, standard deviation, and cofidence intervals for a batch of 10 standard estate samples (SES).

Thus, if the leaf analysis result for a particular block was 2.3 %N, based on the SES analysis the true value lies between 2.3 +/- 0.06, or 2.24–2.36. If the critical N% level is 2.36 we must conclude that Block 23 is not deficient.
Similarly, if the leaf analysis result for a particular block was 1.3 %K, based on the SES analysis the true value lies between 0.9 +/- 0.042 or 0.858-0.942. If the critical K% level is 0.95 we may conclude confidently that Block 23 is deficient.
In the above example, the variability for the test samples is rather small.

No comments: