Rotation

Rotation background

Modelling of cropping or crop-pasture rotations to date has primarily been based on a predetermined restricted set of rotations represented as “activities” in a LP matrix [WE00]. However, this approach often limits the potential rotations that can be selected by the model and does not capture the flexible nature of real-life rotation selection especially in the face of unfolding seasonal conditions. For example, using a fixed rotation structure, it is not possible to alter the rotation in response to the timing of early season rainfall. It also results in the necessity to build entirely new modules for each agro-climatic region due to differences in crop and rotation choices that are available and applicable to each region.

In AFO, we adopt an alternative method proposed by Wimalasuriya and Eigenraam [WE00], where the “activities” in the model are rotation phases. A rotation phase is a land use with a specific sequence of prior land uses [1] (‘history required’). A constraint is included to ensure that for the model to select a given rotation phase, the ‘history required’ must match the ‘history provided’ from another rotation phase. The model solves for the optimal rotation through a selection of rotation phases. This is an unrestricted approach that supports a large range of possible rotations and allows greater flexibility for adding new land uses. Additionally, the approach aligns closer to reality, facilitating a more detailed and accurate representation of the effects of weather-year type on rotation choice.

Each rotation phase requires a history and provides a history. As a simple example, consider the rotation phase barley - wheat: canola in which canola is the current land use. Barley followed by wheat is the history required and wheat followed by canola is the history provided. Based on the current land use and the land use history the level of production (grain and stubble production from crop phases and, seed set and germination from pasture phases), the costs, the machinery requirement and the labour requirement are determined.

The terminology used in AFO to distinguish each land use in a rotation is that the current land use is termed year 0. The prior year land use is year 1 and so on working backwards through the rotation phase. For example, in the rotation Canola: Barley: Wheat; Canola is year 0, Barley is year 1 and Wheat is year 2.

The rotation phases are designed to be as simple and general as possible while still capturing important performance and management variants. The system employed is to generate all possible combinations of the land use sequences over a set number of years, then the infeasible options are removed and the remainder are generalised where possible e.g. wheat, barley and oats may be generalised to cereal in the phase history if the type of crop does not affect subsequent productivity or costs.

The length of the rotation phases and the level of generalisation is determined so that the impacts of the history on the current land use production and costs are captured. These can be summarised by:

  1. The need to track the number of crop phases to determine if an annual pasture needs reseeding.

  2. The need to track the effect of a land use on the productivity or costs of subsequent land uses. This can be either:

    1. Fixing of soil nitrogen and its subsequent effect on following crops. This requires tracking:

      • The number of years of the legume as it affects the quantity of organic nitrogen.

      • The number of years since the legume to determine the remaining nitrogen.

    2. Impacts on disease levels

    3. Impact on weed seed levels

  3. The impact of cropping on subsequent annual pasture seed bank and germination.

The impacts and assumptions of land use history on production and costs that are being captured in the rotation phases developed are:

  1. Annual pasture will be resown if the four most recent land uses in the history are crops. Resowing impacts the current year and the succeeding year.

  2. Lucerne (or Tedera) will be resown if the immediately preceding land use is not Lucerne (or Tedera).

  3. The impacts of spray-topping and manipulating pastures lasts for two years.

  4. Germination of annual pasture is affected by:

    1. The two most recent land uses in the history.

    2. The crop type immediately prior to the annual pasture. Specifically:

      • Germination is higher after an oat fodder crop.

      • A pulse crop increases growth of annual pastures (which is represented by an increase in germination)

  5. A history of legume pasture (annual, Lucerne and Tedera) provides organic nitrogen for subsequent non-legume crops (cereal or Canola).

    1. The amount of organic nitrogen increases up to four years of consecutive legume pasture.

    2. The impact of the organic nitrogen lasts for a maximum of three years.

  6. Pulse crops provide organic nitrogen for subsequent non-legume crops.

    1. The impact of the organic nitrogen lasts for a maximum of three years.

  7. Leaf disease and root disease builds up for each land use and reduces productivity for consecutive land uses.

    1. It is assumed that the maximum level of disease is reached after 4 consecutive years of a land use.

    2. There is variation in the length of the break (interval in years between the same land use) required and the duration of the benefits of a break.

To capture all the factors listed above, the length of the rotation phases represented in AFO is defined to a maximum of 6 years, allowing a history of 5 pastures to be tracked. To reduce the number of rotation phases, land uses in the history that are assumed to have the same impact on the production and cost of the current land use are grouped into ‘land use sets’ (see Table 2). Whilst still capturing the factors listed above, the following generalisations apply to history in rotations:

  1. All cereal crops can be treated the same, except for fodder which must be tracked in year 1.

  2. All pulses can be treated the same.

  3. All crops can be treated the same after year 3.

  4. Resown annuals can be treated as any annual pasture after year 1.

  5. Manipulated and spray-topped pastures can be treated as any pasture after year 2.

After accounting for these generalisations, the following land use options as listed in Table 1 are represented in each year of a rotation phase:

  • Year 0: all the land use options need to be included

  • Year 1: E1, N, OF, P, A, S, U, X, T, J, ar, sr, m

  • Year 2: E, N, P, A, S, M, U, X, T, J

  • Year 3: the sets E, N, P, A, U, T

  • Year 4 → year X [2] : the land use sets Y, A, U, T

Some of the rotation phases constructed will be illogical and are removed. For example, annual pasture is only resown after 4 years of continuous crop therefore any rotation phase that are generated with resown annual that do not have 4 years of crops preceding it can be removed. See RotGeneration for the full list of illogical rules.

To further reduce the possible number of rotation phases in the model, unprofitable and unused land sequences are removed. See RotGeneration for the full list of rules.

Table 1: Land uses represented by the rotation phases.

Key

Land use

a

Annual (no chemical)

ar

Annual resown

b

Barley

bd

Dry sown Barley

f

Faba

h

Hay

i

Lentil

j

Tedera (manipulated)

jc

Continuous Tedera (manipulated)

jr

Tedera resown (manipulated)

k

Chickpea

l

Lupins

m

Annual (manipulated in winter)

o

Oats

od

Dry sown Oats

of

Oats fodder crop

r

Canola (RoundUp Ready)

rd

Dry sown Canola (RoundUp Ready)

s

Annual (spray-topped)

sp

Salt land pasture (saltbush plus pasture understory)

sr

Annual (spray-topped & resown)

t

Tedera (mixed sward)

tc

Continuous Tedera (resown every 10yrs) [3]

tr

Tedera resown (mixed sward)

u

Lucerne (mixed sward)

uc

Continuous Lucerne (mixed sward)

ur

Lucerne resown (mixed sward)

v

Vetch

w

Wheat

wd

Dry sown Wheat

x

Lucerne (monoculture)

xc

Continuous Lucerne (monoculture)

xr

Lucerne resown (monoculture)

z

Canola (Triazine Tolerant)

zd

Dry sown Canola (Triazine Tolerant)
Table 2: Land use sets used in the rotation phase histories.

Key

Land use set

A

Annual (a, ar)

E

Cereal (b, h, o, of, w)

E1

Cereal without fodder (b, h, o, w)

J

Manipulated Tedera (j, jr)

N

Canola (r, z)

P

Pulse (f, i, k, l, v)

S

Spray-topped annual pasture (s, sr)

T

Tedera (t, tr)

U

Lucerne (u, ur)

X

Manipulated Lucerne (x, xr)

Y

Anything not annual (E, E1, P, N, J, T, U, X)

Land heterogeneity

Land quality can vary significantly across a farm, predominantly due to variations in soil type. This can impact:

  1. The efficiency of machinery use and its rate of wear.

  2. The proportion of arable area.

  3. The level of inputs applied.

  4. The level of production.

AFO represents land variation by splitting the farm area into a specified number of land management units. Each land management unit has its own parameters reflecting the factors as listed above. Phases of rotations are possible on each land management unit. The model can solve how to best utilize the area of each land management unit. Each LMU has a specified proportion of arable area. Non arable area can not be cropped however it is accessible by livestock.

Note

If different crop management prior to a pasture phase (e.g. reduced expenditure on herbicide because weed seed control is not important) is to be represented then this will require extra landuse options for the previous crop and extra landuse options for the germinating pasture.

Note

The FOO level at the end of spring is not carried into the next year as a weed burden (e.g. the weed burden in the current year is the same independent of grazing/season last year). This could be represented by adding pasture phases that have spring FOO level as a part of the definition e.g. EEEEah which is annual pasture with high spring FOO following multiple cereals.

Rotation generation

Precalcs

Pyomo