The most extensive data on nutrient requirements exists for vegetables. The reasons likely include:

The table below provides the total nutrient requirement for the main vegetable crops grown in open fields in Germany. This table is particularly useful because it:

This makes it applicable to a wide range of information scenarios. However, I am always bothered by the presentation of P, K, Ca, Mg, etc., in their oxide forms. We know that these often do not exist in the soil in this form, and they are certainly not taken up by plants in this form.

It is common practice to estimate nutrient requirements using "average yields" or "target yields" from multiple farms, a region, or an advisory group. However, the following figures illustrate that this approach can lead to significant errors. The following diagrams show how widely white cabbage yields can vary between different farms.

To determine the fertilization rate, knowledge of the crop's total nutrient requirement is essential. However, it is also interesting and important to know what proportion of the total plant constitutes the harvested product (yield). This significantly influences what portion of the nutrient requirement is removed from the field with the harvested product. The figure below shows the marketable yield and, expressed relative to it, the proportion of harvest residues. The range of both values is enormous and depends, among other factors, on which plant part serves as the yield (see also figures in the "Total Requirement" section).

From the above, it follows that the proportion of nutrients taken up and contained in the yield can also be expected to vary greatly among different vegetable species. This is because the optimal potassium content of a plant varies far less than yields do (potassium content ranges from 20–40 g/kg dry matter, while yields range from 80–1,000 dt/ha). For this reason, it is useful to examine nutrient removals for different vegetable species when normalized to the same marketable yield. This provides insight into their "nutrient demand" and, potentially, the expected yield response to fertilization.

The underlying philosophy for estimating nutrient requirement based on yield is the assumption of proportionality: if you know the yield, you can calculate the requirement (or the nutrient removal with the harvested product). One might be tempted to use the nutrient concentration in the plant (if it is easy to measure or measured routinely) to determine nutrient removal.

The figures below show that for red cabbage (approx. 50% harvest residues, approx. 50% marketable yield), the first assumption holds true, but the second does not. This is surprising, as nutrient concentration is indeed part of the calculation for nutrient removal. So, what is the reason for this discrepancy?

Greenhouse production is generally characterized, compared to open-field cultivation, by:

Therefore, data on nutrient requirements, removals, and uptake are often expressed per square meter. Converting from g/m² to kg/ha is easily done by multiplying by a factor of 10 (e.g., 35 g/m² = 350 kg/ha).

The tables below show the nutrient removals (or nutrient amounts taken up) for various crops, planting densities, and yield levels.

When considering these figures, one should be aware that the achievable yields in a greenhouse depend heavily on the available technology and the input of production factors. Yields of 50 kg of tomatoes/m² are entirely possible. The distribution of absorbed nutrients in tomato and cucumber plants is shown in the two tables below. 60-  80% of the applied nutrients are found in the harvested products.