|
Why Monitor Seasonal Pastoral Conditions? (by
Col Paull, March 2002)
The variability of the environment in pastoral areas of
Australia affects pasture production, animal production, carrying capacity and
profits. The potential for land and pasture degradation means that both governments
and industry have a responsibility to monitor the condition of these resources,
and to derive sustainable management practices which account for year-to-year
variation in climate and pasture production.
Pasture research over the last 30 years, combined with modern technologies,
is opening up new possibilities for improving pasture management decisions.
If we are to manage our grasslands and rangelands profitably and sustainably,
we need to have access to the best up-to-date, big-picture information on seasonal
conditions. This provides an important part of the context within which graziers
make property management decisions.
Assessing Pastures
Pasture studies have indicated that adverse changes in pasture composition are
determined by the degree of pasture utilisation; land degradation is determined
by the amount of surface cover which is closely related to standing pasture
dry matter. Stocking rate affects the processes of pasture growth, detachment
and animal composition.
Thus quantification of the interaction of stocking numbers and pasture growth
is required in order to understand processes, define sound management practices,
monitor resource use and forecast the risk of degradation.
It is on this basis that a spatial modelling framework was developed for Queensland
to provide up-to-date simulations and forecasts of pasture growth and utilisation.
The Project
In the first stage of the project, completed in 1996, work
in Queensland successfully integrating climatic analysis, natural resource data,
remote sensing, historical agronomic research and simulation modelling (to provide
regular assessments of drought) was treated as a case study. While seasonal
rainfall explained only 40% of the variation in seasonal pasture growth, models
of soil water and pasture growth could explain 40-70% of observed variation.
The approach involved running a pasture growth model for tropical and sub-tropical
grasses (GRASP) on a 5 km grid over the entire State, and calibration of it
for a broad range of pasture communities, soil types and climatic conditions.
While prototype simulations of pasture growth over the whole continent were
demonstrated, the need for regionally specific models was indicated.
The second stage of the project, called Aussie GRASS, was commenced in April
1997. It involves nationwide collaboration, in both research and extension,
to operationalise the spatial modelling for assessing the condition of Australia's
grazing lands.
Rainfall Maps
Production and Interpretation
Two types of rainfall maps are available on The Long Paddock
website: total rainfall maps, and maps of 'rainfall relative to historic records'
which are presented as percentiles.
Average and Median Rainfall
Average (or mean) and median are two 'measures of centre' of a set of observations.
The average rainfall for a particular month is the sum of all total rainfalls
for that particular month divided by the number of years of records. Median
rainfall is calculated by ranking, from lowest to highest, the total rainfall
measurements for the period in question. The median is the rainfall total positioned
in the middle of the ranking. When comparing the total rainfall for a period,
to historical records, the median rainfall gives a better indication of 'usual
rainfall', than the average rainfall does (particularly for inland areas).
Rainfall Percentiles
Rainfall percentiles are an extension of median rainfall. The lowest 1% of the
ranked rainfalls is the first (1st) percentile, the next highest 1% is percentile
range 2, and so on, the highest 1% of the ranking being percentile range 100.
The percentile ranges shown on these maps give an indication of how dry or wet
a period has been, relative to the historical record.
Accuracy of Maps
These maps are produced from rainfall reported at points throughout the State
(to the Bureau of Meteorology), and their accuracy varies with the density of
rainfall stations. For example, they are most accurate in areas of high population
density, and least accurate in sparsely-populated pastoral areas. Thus the accuracy
of the spatial maps at a particular locality can be improved by increasing the
number of rainfall stations reporting measurements to the Bureau of Meteorology
from that locality.
Rainfall measurements at point locations are checked for errors, and missing
data calculated using interpolation techniques. The same techniques are then
used to estimate the rainfall received in areas between the recording stations.
Pasture Maps
Production and Interpretation
Rainfall maps give us some idea of pasture production.
However, the effectiveness of total rainfall over a particular period depends
upon previous seasonal conditions, rainfall distribution, rainfall intensity,
temperatures, soil and pasture types, and time of the year. Thus we produce
a pasture growth percentile map, similar to rainfall percentile maps, which
provides a more accurate picture of seasonal conditions.
In order to interpret a map of 'Pasture Growth Relative to Average', it is necessary
to know in outline how the map is produced and its accuracy.
Map Production
Using the Aussie Grass spatial pasture growth model, we can estimate the total
amount of pasture growth over any consecutive 12-month period. Thus, the most
recent of these periods can be compared to the totals for all previous years
covering the same months of the year. A useful way to do this comparison is
by using pasture growth percentiles.
The percentile ranges shown on these maps give an indication, for each location,
of how good or how bad the pasture production has been over the nominated period,
relative to the historical record. That is, the map puts the amount of growth
at a particular location into the context of what is 'normal' or 'reasonable'.
A CRAY super computer is used to run the model for each of 70 000 pixels covering
Queensland, and 250 000 throughout Australia. Each pixel covers an area measuring
5 km by 5 km. The modelled growth is an estimate of the effectiveness of rainfall,
allowing for what is known about soils and pasture species in each pixel, and
the effects of temperature and humidity over the period. The climatic values
used for a particular pixel are derived from grided data layers produced from
corrected point data using interpolation techniques.
Note that the 'Total Pasture Growth (kg of dry matter/ha)'map shows calculated
pasture growth, and is an assessment of how good the seasonal conditions have
been for growth. As it does not take into account feed carried over from the
previous year, or feed eaten, it may not reflect current feed reserves.
On the other hand, the maps entitled 'Total Standing Dry Matter (kg DM/ha)'
show the calculated amount of dried plant material present. The calculation
takes into account grass growth, detachment of plant material from grass plants,
animal intake and the carry-over of standing plant material from previous seasons.
Thus the scope for inaccuracies in the calculation is increased.
Accuracy
These products are designed to be used to provide information
on areas down to a size of one-quarter to one-half of a local government area
(for example, a Shire in Queensland). This means that the user of the products
should look for useful accuracy down to a locality level rather than at the
property level, that is areas covering about 100-250 square kilometres (4-10
pixels). You are less likely to misuse the scale of a mapping product if you
use an A4-sized map of a State or Territory rather than an enlarged version.
As the model is experimental, the accuracy of this information product has yet
to be validated for areas outside Queensland. Field data is currently being
collected to allow the model to be fully calibrated and validated. The map may
contain inaccuracies in some localities, particularly due to the sparse rainfall
reporting network in certain areas of the country.
The most reliable product to use, at this stage, is 'Pasture Growth Relative
to the Last 40 Years'. Other products from running the GRASP model generally
require some further validation.
Please advise us if you find significant problems, and we will attempt to improve
the inputs to the model and its processes.
Use in Making Decisions
The map of 'Pasture Growth Relative to Average' supplements
and supports the information contained in some of our other mapping information
products. For example, in October 1996 calculated pasture growth in the below-average
to extremely-low range corresponded well with areas in Queensland which were
declared drought-stricken, and with the map of 'Rainfall Relative to Historical
Records' covering the previous two years.
In addition, calculated pasture growth in the above-average to extremely-high
categories corresponded well with the areas which had received above-average
to extremely-high rainfall over the previous 12 months.
Thus the pasture growth map can be used in conjunction with rainfall maps, and
also for example the Queensland Drought Situation Map, to assist with decisions
involving feed budgeting, estimating carrying capacities, the most likely areas
for agistment, the use of fire as a pasture management tool and stock prices.
|
