SupFeed module

author: young

Supplementary feeding is the supply of additional feed to livestock, primarily grain and hay (which are both represented in the model). Supplementary feeding is commonly used to help meet production targets such as lamb growth rates prior to sale, or to fill the feed gap to allow higher stocking rates during the summer and autumn months when pastures and crop residues are limiting. Additionally, feeding supplement can be used as a tactic to allow pastures to be deferred early in the growing season which increases subsequent pasture growth rates through increasing leaf area index.

AFO represents a range of supplements including, oats, lupins and hay. Grain and hay as supplementary feeds can either be grown on farm or purchased from another farmer at a farm-gate price (i.e. net price of a product after selling costs have been subtracted) plus the transaction and transport costs. Supplementary feeding incurs a depreciation cost associated with storage infrastructure and variable costs associated with insurance, silo preparation, insect management, grain shrinkage/losses and machinery usage when feeding the supplement. The costs are calculated per tonne to allow for variations in grain density and amounts fed. Supplementary feeding also incurs a labour requirement for time spent traveling to and from the silo, filling the sheep feeder, emptying the feeder, and transporting between paddocks.

Note

Other grains can be added as supplements. Just remember to add their inputs in the mach sheet in universal.xlsx for each machine option and the sup sheet in both universal.xlsx and property.xlsx.

AfoLogic.SupFeed.f_buy_grain_price(r_vals)

Cost to purchase a tonne of supplement off farm.

Purchase price of grain off farm is slightly different to the selling price. The purchase price is equal to the price the selling farmer would receive had they sold to market (ie farm gate price) plus a transaction fee and transport cost.

AfoLogic.SupFeed.f_sup_cost(r_vals, nv)

Machinery, storage and depreciation costs incurred to feed 1t of supplement.

Grain for supplementary feeding is generally stored in large on farm silos and hay in a hay shed. There are both variable cost and depreciation costs associated with the storage of supplementary feed. The variable cost represents the expenditure on insurance, silo preparation, insect management and grain shrinkage/loss. The depreciation represents the yearly reduction in the value of the silos and hay sheds.

In the short term storage costs are fixed, that is, the farmer incurs the same variable and depreciation cost of storage independent of the amount of supplement fed. Additionally, the farmer is limited to only feed as much supplement as the storage capacity. However, AFO is built to evaluate the medium term where storage capacity can be varied. To account for this the cost of the storage is divided by the storage capacity returning the storage cost per tonne of supplement. This cost then applies to each tonne of supplement fed. The storage cost is calculated assuming that all supplement fed for the year is stored on farm. If the capacity of the silos is less than the supplement fed (meaning additional supplement is purchased part way through the year) then the storage cost will be overestimated. However, discussions with farm consultants suggest that farmers store enough supplement for the whole year and extra to handle a poor year.

The machinery cost to feed a tonne of supplement is added in this function however it is calculated in Mach.py (see Mach.py for details on machinery cost to feed supplement).

AfoLogic.SupFeed.f_sup_labour(nv)

The labour required to feed sheep one tonne of supplement is calculated as the time spent traveling to and from the silo, filling the sheep feeder, emptying the feeder, and transporting between paddocks. The transport time to and from the silo and the time to fill up are inputs which are divided by the capacity of the feeder to return the hours per tonne. The rate of emptying the feeder at eight different times of the year are inputs which determine the time taken to feed a tonne of supplement at different times of the year. The time spent traveling between paddocks (not to the silo) is calculated based on the estimated feed rate per sheep (g/hd/d), the estimated number of sheep in each paddock and the average time taken to travel to the next paddock.

To improve the accuracy between different grains, the time taken to fill and empty the feeder rate is calculated in cubic meter units (because a m3 is the same for all grain) and then converted to hr/tonne in the last step. However m3/hr is not an input many can relate to, so in the effort of making the model easier to calibrate, the inputs are entered in a more common format for a specific grain and then converted to m3. For example, the inputs used to determine time to empty the feeder are the feeding rate of lupins in kg/sec, this is then adjusted to m3/hr using the density of lupins. The time spent traveling between paddocks (not to the silo) is calculated slightly differently. Driving from one paddock to the next takes a given amount of time. This time is then allocated to each megajoule being fed. This time is associated with the energy fed because farmers have a target for how many megajoules to feed before going to the next paddock. Using energy also allows all grains to be compared on an equal playing field, because the target is the same for each feed.

Note

The main limitations to this method are if the estimated rates of feeding are wrong compared to the model solution, it is a difficult limitation to avoid because the inputs are not aligned with the activities ie there is no activity that is the rate of feeding sheep so we have to make an estimated link between feeding rate and the total tonnes feed.

AfoLogic.SupFeed.f_sup_md_vol(r_vals, nv)

M/D and DM content of each supplement are known inputs. Unlike stubble and pasture, the quantity of supplementary feed consumed (the decision variables) are expressed including moisture content (i.e. as fed). Therefore M/D must be adjusted by the DM content of the feed. The volume of supplementary feed is calculated based on the quality of the feed. So, lower quality supplements (like oats) will substitute more for pasture than high quality supplements (like lupins). In the generator it is assumed that the availability of supplementary feed is high and that supplement is consumed as the first component of the animals diet. Furthermore, it is assumed that if sufficient levels of supplement are offered then all the sheep’s diet will be provided by supplementary feeding. In pyomo, the optimisation can select a combination of paddock feed and supplement that meets the volume constraint while providing sufficient ME intake for the animals. The volume required for supplement is adjusted (to 80%) because this is necessary to make the substitution rate in the matrix (using volumes) align with the substitution rate calculated using the GrazPlan selection routine. The current representation does not include an effect of high protein supplements (like lupins) overcoming a protein deficiency, and therefore acting as a ‘true’ supplement and increasing intake. If this was represented it would likely make low rate lupin supplementation optimal in early summer/autumn to overcome a protein deficiency.

A portion of supplement fed in the paddock is wasted and hence not consumed. This is accounted for by reducing the ME and Vol of the supplement decision variable (the supplement decision variable is the supplement fed not the supplement consumed by livestock). In confinement, it is assumed that no supplement is wasted.

Note

Supplement M/D does not go through f_effective_mei because the quantity of Sup feed can be controlled so the animals achieve their target weight profile and aren’t gaining then losing weight.

AfoLogic.SupFeed.f_sup_params(params, r_vals, nv)