The CLUES Project: Predicting the Effects of Land-use on Water Quality – Stage II

Objective 4a: Enterprise-scale modelling (AgResearch)

OVERSEER® scenario development

A software component for the OVERSEER® nutrient budget model was provided to NIWA (in the form of a Dynamic Linked Library – DLL). Documentation was provided to explain how to call the software, and it was then linked into the CLUES modelling system, as demonstrated in Section 6. The OVERSEER® DLL supplied to NIWA has capability for five farming scenarios:

Dairy.

Sheep/beef lowland.

Sheep/beef hill country.

Sheep/beef high country.

Deer.

For each of the five scenarios, optional stocking numbers may also be supplied. To adequately specify input to OVERSEER®, many other inputs are also needed. Many of these have been set at values that are typical for the farm type and region, making use of MAF Monitor Farm model (MAF 2004 see section 9.7) and other data sources that are described below.

The assignment of MAF Monitor Farm models to regions was done as shown in Table 9-1, plus the following assignments:

The “intensive” models in Table 9-1 were assigned to Sheep/beef lowland.

The “hill” models in Table 9-1 were assigned to Sheep/beef hill country.

Merino model values were assigned to Sheep/beef high country.

Dairy data was obtained from the summary data in Livestock Improvement Corporation (2004), rather than MAF Monitor Farm dairy models

The MAF Monitor Farm Otago Dry Hill model was not used at this stage.

Table 9-1: Assignment of MAF Monitor Farm models for Sheep and Beef, and Deer, to regional council regions.

(Wide table viewable here)

The OVERSEER® DLL used in CLUES assumes that inputs are within a valid or reasonable range, and that there is internal consistency between inputs e.g., fertiliser inputs and productivity.

The other OVERSEER® inputs which can vary within the CLUES framework are:

Rainfall (annual average rainfall in mm).

Region (select one of 15 regional council regions).

Soil order (one of up to 13 possible soil orders).

Topography (one of possible classes for the block slope).

Maps are available within CLUES, which provide the information necessary to automatically estimate each of these 4 other input variables, anywhere in New Zealand.

Calling the OVERSEER® DLL

The standard call for the DLL is:

CluesOvr(scenario, region, soilorder, Topography, rainfall, Sdairy, Ssheep, Sbeef, Sdeer).

The model returns N (usually as nitrate) and P losses associated with each land use.

The definition of each input is:

Scenario: 1–5 for pre-defined scenarios dairy, sheep/beef (lowland), sheep/beef (hill), sheep/beef (high), deer.

Region: Regional council number code (see Table 9-2 for definitions). This code is used to set default regional values for dairy and sheep/beef farms, respectively as shown in Table 9-2 and Table 9-3.

Soilorder: Soil order specified as an integer, based on the following code:

Code Soil order
1 Allophanic
2 Brown
3 Granular
4 Gley
5 Melanic
6 Organic
7 Oxidic
8 Pallic
9 Podzols
10 Pumice
11 Recent
12 Semiarid
13 Ultic

Slope: average block/subcatchment topography code based on the following table:

Code Slope Class Access Slope LRI1 class
1 Flat   0° to 7° A-B
2 Rolling Area mostly navigable by tractor 8° to 15° C
3 Easy >50% area navigable by tractor 16° to 25° D-E
4 Steep <50% area navigable by tractor 26° or more F-G


1
LRI = Land Resource Inventory slope class

Rainfall: Average annual rainfall to nearest 100 mm

Sdairy, Ssheep, Sbeef, Sdeer: Optional stocking numbers for dairy, sheep, beef and deer, respectively. If a zero value is provided then a default is used as described in Section 0.

Region default values are based on latest published set of data from Livestock Improvement Corporation (Livestock Improvement Corporation 2004) and MAF Monitor Farms (MAF 2004). These values do change over time due to variations in economics, farm practices and weather, for example. A sensitivity analysis could be done in future to look at the change over time. At this stage, the important factor is that the model has the structure to hold these values. They are relatively easy to change if the basis for calculation was to change (e.g., using a different base year, or using an average over several years, or using a different published set of data). The base for these values may be something that the users would like to have established as a future step in the project. These default values could be updated either within the DLL (in which case a maintenance agreement will be needed) or the DLL needs to be modified so that external data source containing the default values can be accessed.

Note that in Table 9-3 the nitrate-N values are estimated typical (as opposed to average) annual applications obtained from a fertiliser company. So, for example, on Manawatu/Wanganui sheep/beef hill country farms, there are some farmers using nitrate-N in this region, but typically, the application rate is zero. This data was obtained so that the DLL could work with the GIS interface. In theory, the nitrate-N fertiliser rate should be commensurate with the production data used in the DLL, and hence a more reliable source of information of typical nitrate-N fertiliser rates is required.

Table 9-2: Regional categories and associated default values for dairy farms. Animal production data is based on Livestock Improvement Corporation (2004).

Code Region Average milk production (kg milksolids/ha/yr) Average cows (cows/ha) Average milk per cow (kg milksolids/ha/yr) N fertiliser (kg N/ha/yr)
1 Northland RC 730 2.33 311 80
2 Auckland RC 1060 2.85 372 200
3 Env Waikato 853 2.67 318 120
4 Env. Bay of Plenty 909 2.80 324 120
5 Gisborne 832 2.56 333 70
6 Hawkes Bay 840 2.62 322 120
7 Taranaki 734 2.54 289 100
8 Manawatu/Wanganui 790 2.48 316 50
9 Wellington 604 2.11 286 80
10 Marlborough 910 2.60 348 100
11 Tasman 885 2.60 341 100
12 West Coast 805 2.71 296 150
13 Canterbury 790 2.49 316 80
14 Otago 734 2.54 289 100
15 Southland 635 2.03 316 150

Table 9-3: Regional categories and associated default values for sheep/beef farms. Animal production data is based on corresponding MAF Monitor Farms in Table 9-1.

Code Region Sheep (lowland) Sheep (hill country)
Sheep
(SU)
Cattle
(SU)
Wool
(kg /SU/yr)
N fertiliser
(kg N/ha/yr)
Sheep
(SU)
Cattle
(SU)
Wool
(kg /SU/yr)
N fertiliser
(kg N/ha/yr)
1 Northland RC 5.17 6.43 4.94 0 6.29 3.71 4.76 30
2 Auckland RC 8.67 2.13 4.8 40 3.2 3.2 4.52 50
3 Env Waikato 5.17 6.43 4.94 0 6.29 3.71 4.76 50
4 Env. Bay of Plenty 5.17 6.43 4.94 30 6.29 3.71 4.76 40
5 Gisborne 6.53 5.86 5.4 40 4.95 3.85 4.96 30
6 Hawkes Bay 6.53 5.86 5.4 60 6.69 3.41 4.6 40
7 Taranaki 6.53 5.86 5.4 50 6.29 3.71 4.76 30
8 Manawatu / Wanganui 8.67 2.13 4.8 30 3.2 3.2 4.52 0
9 Wellington 5.17 6.43 4.94 30 4.23 5.57 5.3 20
10 Marlborough 12.78 0.72 5.74 40 7.48 1.22 4.97 0
11 Tasman 12.78 0.72 5.74 40 7.48 1.22 4.97 30
12 West Coast 5.17 6.43 4.94 0 6.29 3.71 4.76 20
13 Canterbury 8.67 2.13 4.8 0 3.2 3.2 4.52 30
14 Otago 5.17 6.43 4.94 50 6.29 3.71 4.76 30
15 Southland 5.17 6.43 4.94 30 6.29 3.71 4.76 30

Operation of OVERSEER® DLL

The CLUES system currently calls OVERSEER® with stock rates developed by Hunt (2003), and given by Hunt as his Table 2. The stocking rates from Hunt (2003) depend on slope class (for which CLUES has a detailed map), location (whether North Island or South Island), and livestock category (dairy, sheep, beef, or deer). In future it may be possible to use the stocking rates developed as part of the land use classification reported in Section 16.

For dairy farms

If a scenario is entered, and stock number are all zero then production is calculated as:

Cow numbers = Average regional cow numbers (see Table 9-2)

Milk production = Average regional milk production (see Table 9-2)

If dairy cow numbers are entered (this is the option CLUES uses) then

Cow numbers = entered value

Milk production = cow number x average regional per cow production (see Table 9-2)

It is assumed that effluent is applied as spray irrigation at an application rate of 150 kg TN/ha/yr of effluent, and that nitrate-N fertiliser on the effluent block is reduced.

For sheep beef farms

If a scenario is entered, and stock numbers are all zero then stock units (SU) are calculated as:

For sheep/beef lowland and hill country

Use SU sheep and SU beef values shown in Table 9-3

Wool = (kg/SU/year in Table 9-3) x SU sheep

For high country sheep/beef

SU sheep = 1.3

SU beef = 0.2

Wool = 4.7 x SU sheep

If stock numbers are entered then

SU sheep = Stock number sheep x 1.1

SU beef = Stock number beef x 5

Note that this option to enter stock numbers is in addition to the requirements of the contract, and values used are only approximate. Fully referenced values can be inserted as part of year 3 of the CLUES project.

For deer

If a scenario is entered, and stock numbers are all zero then SU is calculated as:

SU deer = 12.9 for North Island

SU deer = 13.0 for South Island

If stock numbers are entered then

SU deer = Stock number sheep x 2

N fertiliser rate is a regional estimate, based using the values for sheep/beef (lowland) if topography is flat, otherwise Sheep/beef (hill).

Data sources for OVERSEER® component of CLUES

The data sources for OVERSEER® are summarised in Table 9-4.

Table 9-4: Sources of data for OVERSEER® component of CLUES

Data Description Source of Data Date Expected timing of next update How to obtain updated data
Stock data MAF Monitor Farm reports July 2004 As required, intervals of 6 to 18 months Obtain new OVERSEER® DLL for use with CLUES or have an external data source which the DLL can access
Dairy data Dairy statistics from Livestock Improvement Corporation 2003/4
Fertiliser data Estimate from company reps  

In addition, the underlying model is updated at 12-18 month intervals, and the OVERSEER® DLL should also be updated at the same time.

Additional work

To improve the function of the DLL in its current construct it is recommended the following activities occur:

Upgrade the underlying model to the new version of OVERSEER® nutrient budget model

Improve the structure of the DLL to improve functionality of the link with the GIS system

Develop routines to better check for consistency of data

Proposed next steps

Upgrade the underlying model to the new version of OVERSEER® nutrient budget model

Improve the structure of DLL to improve functionality of the link with the GIS system

Develop routines to better check for consistency of data

Make visible the default values associated with the 5 current scenarios (tentative)

References

Livestock Improvement Corporation (2004). Dairy Statistics 2003-2004, Business Information Unit, Livestock Improvement Corporation Limited, Hamilton. (Available at http://www.lic.co.nz/113_7.cfm).

Hunt, C. (2003). A calculation of stock numbers for four livestock categories on seven classes of slope for New Zealand. AgResearch Report prepared for the Ministry of Agriculture and Forestry, June 2003.

MAF (2004) Sheep and beef monitoring report. Ministry of Agriculture and Fisheries, Wellington (Available at www.maf.govt.nz/mafnet/rural-nz/statistics-and-forecasts/farm-monitoring/

Contact for Enquiries

MAF Information Services
Pastoral House
25 The Terrace
PO Box 2526
Wellington, NEW ZEALAND

Fax: +64 4 894 0721
Contact this person