Source DocumentPrevious PageTable Of Contents

Spatial and temporal trends of CSBP soil and plant analysis data in shires of Western Australia

Andreas Neuhaus1,2, Leisa J Armstrong 2, Sreedhar Nallan2, Keith Ancajas2, Natasha Warnasuriya2 and Aalap Paresh Gandhi2

1 CSBP Limited, Kwinana, Perth, WA 6966 Email
Edith Cowan University, Computer Science, Mt Lawley Campus Email,,,


Nutrient deficiencies in shires of Western Australia (WA) can be determined from soil and plant analysis data to indicate requirements and to generate awareness among advisors and growers. Topsoil analytical results from CSBP’s ASPAC (Australasian Soil and Plant Analysis Council) accredited laboratory have been used to show nutrient limitations. Temporal trends in two contrasting cropping areas showed that soil pH, phosphorus (P) and potassium (K) may reflect historical lime and nutrient applications. More mobile nutrients like nitrogen (N) and sulphur (S) fluctuate seasonally. Unlike in the southern shires, P and K accumulated and pH increased over time in the northern shires. These trends are steady while soil N and S values vary considerably. Main constraints to cereal and canola production in WA are currently low N, K and S (about 30% of samples), P and pH (about 20% of samples). Plant tests indicated mostly adequate copper, zinc and manganese concentrations. Predicting P, K and pH may be possible using trend analysis data, soil property and fertiliser or lime applications on a paddock level. Regular soil sampling, however, is recommended to better understand the dynamics of N and S requirements. A customizable decision support system that visualizes either more accurately estimated or, as in this case, measured nutrient statuses in real-time using geospatial technology at shire level provides useful practical information for advisors.

Key Words

Nutrient deficiency, spatial analysis, nutrient trends, cereals, decision support


Agricultural soils in southwest Western Australia (about 18 M ha) are highly weathered and in general infertile. Continuous fertiliser applications and nutrient removal from crops are changing soil chemical properties. For a long time CSBP emphasised the importance of soil and plant tissue testing for growers to monitor nutrient and pH changes on their paddocks. Fertiliser advice has also been customised to the paddock situation. It is hypothesised that over time soil fertility has improved especially for less mobile macro nutrients while the soil pH may have decreased.


Currently up to 80 000 samples per annum are analysed from WA at the CSBP Soil & Plant Laboratory. Based on this large sample size a comprehensive trend analysis over the last 15 years in two combined cropping areas (northern shires – Northampton, Greenough, Chapman Valley; southern shires – Lake Grace, Kulin, Dumbleyung) has been undertaken. Soil test data from the 2011/2012 season have been mapped for a spatial analysis. Soil data has been selected for 0-10cm depth and for cereals and canola as the planned crop in that year.

Although nutrient deficiencies for N, P, K and S are dependent on many factors such as yield potential, rainfall, crop varieties, soil types, N supply via soil organic nitrogen and residual organic nitrogen and S supply from mineralisation, critical values typical for WA cereal and canola production were used (Brennan and Bolland, 2005, 2006, 2007, Moody, 2007, Robson et al., 1995).

Data analysis

Frequency distributions were used to examine trends over time. Data from 6 shires from 1997 to 2012 were pooled (n = 66 043) into 3 northern and 3 southern shires. Nutrient deficiencies were plotted spatially as number of samples in a shire below the defined critical value.


Following a drought year, the average and standard error of percent of nutrient deficient samples among WA shires in 2011 was 39%2.1% for N, 22%1.6% for S, 20%1.4% for K and 16%1.1% for P. Copper, zinc and manganese deficiencies in 2011 were 9%, 2% and 1% respectively. After the favourable season in 2011, values increased in 2012 for S (33%2.2%), K (32%1.6%) and P (19%1.9%), but declined for N (28%2.3%). Topsoil pH was low in 18% of all samples across WA in 2012.Trace elements deficiencies for wheat were the same as in a drier season such as in 2010 (Cu = 9%, Zn = 2%, Mn = 1%). Low organic carbon levels in the eastern shires point to light soil textures and lower rainfall.

Geographical trends

Figure 1. The spatial distribution of percent of samples below benchmark values in shires of Western Australia. N and S deficiency is predominantly found in northern shires. Soil K and P deficiencies seem to be evenly distributed throughout the WA wheatbelt.

Temporal trends

Figure 2. Bar graphs on the left are trends for the shires of Northampton, Greenough and Chapman Valley. Bar graphs on the right refer to the shires of Lake Grace, Kulin and Dumbleyung. The 5 classes in each chart indicate values for the 1997/1998 to 2011/2012 soil sampling season (from left to right). A drier than usual season prior soil sampling is marked in orange.

Long-term trends of N (not shown), P, K and S (not shown) revealed no correlation with the levels of production from the year before. Phosphorous status appears to have improved greatly over the last 15 years. In the 2010/2011 soil testing season, 75% of samples in the combined northern shires were above the benchmark value. This is a marked increase from the late 1990s when less than 30% of samples were above the benchmark value in those shires. Unlike the northern shires, there are no clear trends in P status in the southern shires. Since 1999/2000, more than half of all results have been at least 40% higher than the benchmark value. This indicates that P application rates have exceeded crop requirements in many parts of WA. However, a low soil pH reduces P availability, especially in shires like Kulin, Lake Grace and Dumbleyung. Potassium concentrations appeared to have improved since 1997/98, but about 40% of samples still have less than 60 mg/kg K in Northampton, Greenough and Chapman Valley.


Soil supply of nutrients has not been build up homogenously. Soil fertility for less mobile macro-nutrients like P improved in the Geraldton, but not in the Lake Grace area. Trends for soil K build up and a pH increase are also more evident for the Geraldton than the Lake Grace area. The changes for mobile nutrients such as N and S are seasonal. This study monitored soil nutrient deficiencies based on many soil samples and a time-consuming data analysis. The data analysis and mapping will be automated to allow the distribution map to appear in real-time as the samples get analysed. Sample volume needs to be high until sensor technologies or predicted values can replace conventional analytical measurements. Sensor technologies are progressing (Bah et al., 2012), which will place more importance on geospatial techniques to analyses and display data.

The results summarised are from samples sent in by CSBP clients and therefore carry that bias. However, a sample size of 50,000 - 75,000 soil samples and about 10,000 - 15,000 plant samples per year probably represents a reasonable resource for estimating the plant and soil profile status of WA soils. Nevertheless, the sample volume per area is about 40 times lower compared with soil tests taken in North America (Fixen et al, 2010). Samples were taken from different paddocks every year and variation for long-term trends of the 6 shires i.e. for the percentage of legume and pasture crops grown in the year before (20-40% for northern and 65-75% for southern shires), percentage of samples with clay texture (1-5 and 5-15%) and percentage of samples with organic carbon above 2% (1-8%) could have partly influenced observed trends. Summer rain events that caused leaching or mineralisation would have caused additional variation in N and S concentrations. Potassium influx from summer rain events could have also slightly impacted the K concentrations.

The assumption that high production in 2011 in northern shires has drawn down soil nutrient concentrations – and N in particular, is confirmed by spatial analysis. Similarly, in areas where there was significant leaching pressure and/or high yields, S reserves are likely to have been depleted, which may explain the % of higher S deficiencies in 2012 compared with the previous year. The soil profile distribution of particularly S and K would be an important additional consideration for the yield response. Subsoil sampling for S and K would give more confidence about their deficiency status.

Much of the State has accumulated good P reserves, and soil test concentrations are more likely to reflect historical P application rates than seasonal effects. P use-efficiency is reduced though, because of problems with soil acidity in many places. And this does not take subsoil acidity into account. The pH situation in WA is much worse than a soil survey described for soils in North America (Fixen et al, 2010).

Plant testing further improves the understanding of trace element requirements. Copper seems to be the most important when compared with Zn and Mn. Although most WA soils have adequate trace elements (Cu, Zn, Mn), deficiencies could increase if soil acidity is not addressed and have the potential to severely limit production.


WA has a good soil analysis, advisory and fertiliser distribution system, but the example of North America shows that more soil testing should be done. North America shows that zone and grid sampling is likely to also lead to an increase in soil testing in WA. Overall topsoil fertility has improved in WA, although pH continues to be a major limitation to crop production.

Here, the reported aggregated values for shires are no substitute for site-specific soil and plant tissue testing. These results are not intended for the use of giving fertiliser advice. They are simply drawing attention to broad nutrient needs of N and S (northern areas), P and lime (throughout the wheatbelt) and K (not limited anymore to only coastal sandy areas) as well as making advisors and growers aware of nutrient changes over time in different locations of WA.

Further practical uses of the trend analysis could be used to improve grower and extension officers strategies for fertilizer management and soil ameliorization, The use of decision support tools that could include geospatial data analysis using gps based data sets could be used by growers to carry out customized scenario planning based on their individual paddock trends as well as for other stakeholders to use these for district and regional planning. Future plans for this research are to develop a software tool that integrates these nutrient data samples and geographic mapping to provide visualizations to growers using data mining and geospatial techniques.


Dr Geof Proudfoot (Manager Agricultural Laboratory, CSBP), Doug Hamilton (Services Coordinator Precision Agriculture, CSBP), and James Easton (Field Research Manager, CSBP) are acknowledged for their valuable contribution to this paper.


Bah A, Balasundram SK and Husni MHA (2012). Sensor Technologies for

Precision Soil Nutrient Management and Monitoring. American Journal of Agricultural and Biological Sciences, 7, 43-49.

Brennan RF and Bolland MDA (2005). Soil and tissue tests for the sulphur requirements of canola. Agribusiness Crop Updates,2005, 11-12.

Brennan RF and Bolland MDA (2006). Soil and tissue tests to predict the potassium requirements of canola in south-western Australia. Australian Journal of Experimental Agriculture, 46, 675 – 679.

Brennan RF and Bolland MDA (2007). Comparing the potassium requirements of canola and wheat. Australian Journal of Agricultural Research, 58, 359 – 366.

Fixen, PE, Bruulsema, TW, Jensen, TL, Mikkelsen, R, Murrell, TS, Phillips, SB, Rund, Q and Stewart, WM (2010). The Fertility of North American Soils. Better Crops, 94 (4), 6-8.

Moody PW (2007) Interpretation of a single-point buffering index for adjusting critical levels of the Colwell soil P test. Australian Journal of Soil Research, 45, 55-62.

Robson, A.D., Osborne, L.D., Snowball, K. and Simmons, W.J. (1995). Assessing sulphur status in lupins and wheat. Australian Journal of Experimental Agriculture, 35, 79-86.

Previous PageTop Of Page