Next Article in Journal
Effects of Soil pH and Fertilizers on Haskap (Lonicera caerulea L.) Vegetative Growth
Next Article in Special Issue
Highly Porous and Nutrients-Rich Biochar Derived from Dairy Cattle Manure and Its Potential for Removal of Cationic Compound from Water
Previous Article in Journal
Toward Precision in Crop Yield Estimation Using Remote Sensing and Optimization Techniques
Previous Article in Special Issue
A Discussion on Mehlich-3 Phosphorus Extraction from the Perspective of Governing Chemical Reactions and Phases: Impact of Soil pH
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Technical Note: Regression Analysis of Proximal Hyperspectral Data to Predict Soil pH and Olsen P

1
School of Agriculture and Environment, Massey University, Private Bag, Palmerston North 4442, New Zealand
2
Ravensdown Fertiliser Ltd., P.O. Box 1049, Christchurch 8042, New Zealand
*
Author to whom correspondence should be addressed.
Submission received: 27 January 2019 / Revised: 8 March 2019 / Accepted: 13 March 2019 / Published: 15 March 2019
(This article belongs to the Special Issue Soil Fertility)

Abstract

:
This work examines two large data sets to demonstrate that hyperspectral proximal devices may be able to measure soil nutrient. One data set has 3189 soil samples from four hill country pastoral farms and the second data set has 883 soil samples taken from a stratified nested grid survey. These were regressed with spectra from a proximal hyperspectral device measured on the same samples. This aim was to obtain wavelengths, which may be proxy indicators for measurements of soil nutrients. Olsen P and pH were regressed with 2150 wave bands between 350 nm and 2500 nm to find wavebands, which were significant indicators. The 100 most significant wavebands for each proxy were used to regress both data sets. The regression equations from the smaller data set were used to predict the values of pH and Olsen P to validate the larger data set. The predictions from the equations from the smaller data set were as good as the regression analyses from the large data set when applied to it. This may mean that, in the future, hyperspectral analysis may be a proxy to soil chemical analysis; or increase the intensity of soil testing by finding markers of fertility cheaply in the field.

1. Introduction

Fertilizer is the largest farm working expense in New Zealand hill country farming systems. Fertilizer applications are usually based on soil test results. This work seeks to establish if it is possible to measure soil fertility parameters using hyperspectral analysis. Large data sets are used to find the wave bands that may be proxies for laboratory testing [1].
New Zealand’s hill country farming systems are largely based on sheep, beef, and sometimes deer animal production grazing on ryegrass (Lolium perenne L.) and white clover (Trifolium repens L.) swards. Easier topography exhibits higher concentrations of both, while harder hill country swards are dominated by species with lower feed value such as brown top (Agrostis capillaries L.) and crested dogstail (Cynosurus cristatus L.) with little clover. Steeper hill slopes tend to have higher concentrations of weed species such as (Cirsium arvense (L.) Scop.) and (Ulex europaeus L.).
New Zealand’s temperate climate allows year round outside grazing. In the most severe environments, such as that created by altitude, animals are typically grazed on lower and better soils during winter and early spring. This is especially the case through lambing and calving in the South Island, and to a lesser extent the North Island’s central plateau.
Sheep follow cattle in a grazing rotation as sheep graze shorter pasture than cattle and each animal class has different classes of parasites and parasitism is reduced as the parasite larvae of each class is destroyed by the grazing of the other. This farming system is heavily dependent on summer rainfall that is quite variable, with higher rainfall on the West coast than the East coast of both islands [2,3], see Figure 1. Where summer rainfall allows grass growth to finish stock farm incomes are higher than on steeper, colder country, where in dry summers farmers are, forced to sell stock, unfinished as store, for others to finish on softer country with irrigation [2]. On the steeper country, the use of vehicular traffic is not possible, and fertilizer and herbicide applications are undertaken by aircraft [4].
Soil sampling for chemical analysis is expensive and limits the number of samples that it is economic to analyze. This results in the averaging of soil samples over large areas up to several hundred hectares for economies of scale [5]. The provision of an extremely large data set of soil samples (the current price for a soil test is around US$70) made it possible to undertake some analysis as this provided the actual data for this work. The cost of undertaking soil testing limits the intensity and is another reason that samples are combined to provide an average. In addition, the samples were measured for spectral reflectance, which provided an opportunity to marry 2150 variables being the spectra reflectance values. This work was undertaken for another project so provided a unique opportunity to initiate work to see if soil elements can be predicted. In addition, a second data set was available at Massey, which had also been collected for other work, was available for testing [6]. The farms that were sampled would not have contemplated such intense soil testing if they were required to pay for it.
At present, on hill country farms, sampling is undertaken in transects that combine samples taken 10 m apart to average values from 27 samples [5]. Combining the samples eliminates the variation, especially for Olsen P. This test is the most important, in hill country fertilizer recommendations. The variation within transects is much greater than the 10 m sampling distance, therefore, there is no geospatial interpolation that is meaningful between sampling points [5,6,7]. Hyperspectral analysis has the potential to be a useful tool for proxy measurements of soil nutrients that could increase sampling density and reduce cost.
Hyperspectral analysis of soil using proximal visual and near infrared (VIS-NIR) spectra have been undertaken for some years. A review of the work indicates that it has been more useful in providing a proxy for soil organic matter and clay content than in predicting elements. However, clay compounds such as kaolinite and minerals such as calcite (a form of calcium carbonate) and pH, have been measured with some degree of success [8]. The proximal hyperspectral sensor used in this work is the Fieldspec 4 ASD (Analytical Spectral Devices, Malvern Panalytical, Malvern, UK), which has also had some success in mining large data sets for soil sensing [9,10,11]. These have shown that in clay, the total organic carbon and moisture have been measured, with variable accuracy, more so in soils with high soil organic matter [11]. Measurement of pH [9] and soil nitrogen [12] has been possible, but has been less robust in predictions.
This work aims to find the spectral wavebands that are proxies for Olsen P and pH, which are the two most important soil analyses used in New Zealand hill country farming. This is because plant available P reduces as soil acidity increases, acidic soil conditions are the norm in New Zealand. Measuring Olsen P using proximal hyperspectral analysis has not been successfully undertaken and pH with limited success.

2. Materials and Methods

2.1. Chemical Analyses

A data set of 3189 soil samples that had been analyzed, for full soil tests at Analytical Research Laboratory an ISO 17025 soil laboratory was married to spectra from the same soil samples collected using a probe on a Fieldspec 4. These measurements were undertaken at the Analytical Research Laboratory in Napier, New Zealand. The samples were from the four farms Ravensdown collected samples shown in Figure 2. These farms are part of a Private/Government research program between Ravensdown Ltd. and the New Zealand Ministry of Primary Industries, which requires extensive soil testing to 7.5 cm [13]. The project is to improve the performance of steep pastoral country and improve the efficiency of fertilizer delivery by aircraft. There is very little cropping and arable production on this land and all the samples are from pasture, which is the focus of the research. This was the original large data set used for the multiple partial least square regression (PLSR) using “R” 3.41 to identify wavebands of interest. Only Olsen P and pH were regressed as the second data set obtained independently had only Olsen P and pH measured at Massey University soil laboratory, which analyzed the 883 soil samples. The smaller Massey data set was sampled at depths (0–3 cm, 3–15 cm, and 15–30 cm). Whereas, all samples from the larger Ravensdown data set were at (0–7.5 cm).
The second data set was from two stratified nested grid samples from different fields on the same farm used to look at geospatial variance with soil depth [6,7]. The samples were tested for Olsen P using the chemical extraction methods and photo-spectrometer was used for the geospatial analysis [6,7]. Those that had not already been measured were completed using the method employed in References [7,14]. The pH measurements were taken using a digital probe in 10 g samples in 25 mL of distilled water. All Fieldspec 4 ASD measurements, at Massey and in Napier were on dry samples to eliminate the effects of moisture [15].

2.2. Statististical Analysis

A PLSR of 2150 wavebands of spectra data from the Fieldspec 4 (350–2500 nm) of 3189 samples were regressed against the chemical analysis results. A PLSR assumes the data to be normally distributed, the mean and standard deviations of the data sets are presented in Table 1. The shape of the distributions are shown in Figure 3a–d.
The Olsen P distributions for both data sets resembled a Chi2 more than a normal distribution. For this reason a log10 of Olsen P of the Ravensdown data and Massey data was regressed as this passed the normality test (see Figure 4).
For this work, the larger data set was regressed using “R” version 3.41 [11,16,17]. The 100 most significant wavebands were selected for Olsen P and pH to regress the smaller Massey University data set. This is because the Massey data set had 883 samples so the number of waveband (factors) to be regressed against as the independent variables should be a much smaller quantity. The regression was undertaken initially using “R”, then repeated using an Excel plug in StatTools version 7.5 (Pallisade, NY, USA). This was undertaken to check that the results were the same.
The regression equations from the Massey University data set were then used to predict Olsen P and pH on the large dataset. The predictions were then compared to the actual laboratory values. This was to establish if the regression of a large data set had found a proxy means of measuring pH and Olsen P using independent data sets.

3. Results

The regression analyses on the data sets for pH and Olsen P, summarized in Table 2, for Ravensdown Ltd. and Massey University data, showed a significant trend between the modeled and actual values for all data set (p value < 0.0001). However, there is a high degree of variability within the data sets shown by the high standard errors and low adjusted R2 values, with the best predictions being for pH. The summary of the scatter plots obtained from the regression are in Figure 5a–f.
The Ravensdown regression analyses have predictions using the regression equations developed on the Massey data. The number of wavebands was reduced from 2150 to 100 to reduce auto correlation and co-linearity effects from the regression, which are inevitable from such a large data set; especially if moving to a smaller data set with less rows than columns used in the PLSR. The significance of waveband contribution were all significant less than 0.05. The regression analyses using R on the Ravensdown data suggest that even though there are more rows than columns regressing using 2150 columns has an auto correlation effect (see Table 2). The equations from the analyses appear in Appendix A.

4. Discussion

The Massey soil data came from 30 cm soil cores sectioned 0–3 cm, 3–15 cm, and 15–30 cm soil depths, which contained samples from the lowest strata with extremely small levels of Olsen P some samples having undetectable levels. This meant that the Massey data regression for Olsen P passes close to the origin. Olsen P is a measure of exchangeable P and is readily available for plant uptake, while the remainder of the P in the soil is in less available forms organic (biomass/animal excrement) and complexed with Al and Fe. The stratification of phosphate in the soil in closely related to the organic matter stratification. The most popular fertilizer applied in New Zealand is single super phosphate (Ca(H2PO4)2) and the next most popular fertilizer supplying P is diammonium phosphate ((NH4)2HPO4) [18,19,20]. Phosphates are not very soluble, which can be problematic in situations where surface run-off is likely [20]. The majority of phosphate tends to remain near the surface of soil horizons through cation exchange with clay minerals, which is why soil samples are taken to 7.5 cm in New Zealand pastures [5,19]. While the hyperspectral data was regressed against Olsen P, the spectra reflectance will be on compounds containing phosphate, similarly to calcium being found in calcite [8].
The regression of the Massey Olsen P data had an adjusted R2 of 0.55 whereas the Ravensdown data produced an adjusted R2 of 0.27. Both data sets had high variation in data fit, but the trends from the analyses are highly significant. The Massey data probably produced a clearer trend as the stratified sample had a larger range of values with increasing values as samples neared the surface [6,7]. The Ravensdown regression improved to an R2 of 0.34 when the log10 of Olsen P was regressed. These results from the Chi2 shape of the raw data became more normalized when the log values were regressed.
The pH data was more normally distributed than the Olsen P data for both data sets. Both data sets were able to be regressed to the hyperspectral data with similar trends, which were highly significant. The R2 of the Massey and Ravensdown data were about 0.42.
The regression equations found from the PLSR of the Massey data were also used to predict Olsen P and pH of the Ravensdown data, as shown in Figure 5c,d,f. This exercise was undertaken to see if the equations developed from one data set were able to provide a reasonable fit with the other data set when using the same wavebands. While the gradients of the equations in Figure 5c,d,f are quite different; they both bisect the data values almost in half. This suggests that the spectral wavebands selected may be transferable to other data sets and that further research and analysis using these wavebands may be worthwhile.
This exercise shows that there is promise in developing proximal hyperspectral sensing of soil nutrients. This may prove to be a cheaper alternative to chemical analysis if it can be undertaken in situ in the field. This could lead to more intensive testing which would be a benefit in delivering the desired fertilizer nutrient to where it is needed through variable rate application, providing more efficient fertilizer utilization.
This study does demonstrate that taking a large data set from a large range of soil types from many locations over thousands of hectares can produce promising results with statistical significance. Other work in this field has been on much smaller data sets applied in localized areas on similar soils. The results of this study are as good as other work and are a step into finding minerals or proxies for fertility.

Author Contributions

M.W. conceived of the research in discussion with M.G. T.K. undertook the chemical analysis at Massey University. P.B., M.G. and A.P. assisted with sample preparation. M.W. provided the Ravensdown Ltd. data set. T.K. undertook the hyperspectral analysis of the Massey data set. M.G. undertook the statistical analysis. M.G. edited the paper in discussion with the co-authors.

Acknowledgments

This study was funded by, Ravensdown Ltd. and Massey University. Ravensdown paid for the preparation and analysis of a data set, whilst Massey University paid for and undertook the analysis of a second independent data set. This research received no external funding.

Conflicts of Interest

The authors declare no conflict of interest.

Appendix A

The regression equations where (×××nm) stands for the reflectance values for the wavebands from the Fieldspec 4. The predicted Olsen P appears on the y axis using the StatTools software.
Olsen P Massey Data = 17.39357943 + 924.99461634 × 515 nm − 1023.05662383 × 516
nm + 839.64318782 × 549 nm − 1901.74299473 × 550 nm + 1137.32505077 × 551 nm –
1574.15784614 × 611 nm − 2656.44564275 × 612 nm − 1236.46471204 × 613 nm +
713.64885867 × 614 nm + 4529.23578307 × 626 nm − 939.78101407 × 627 nm –
629.35692109 × 628 nm + 3612.28340339 × 629 nm + 1915.06008321× 630 nm +
255.93665506 × 662 nm − 7544.07713423 × 663 nm + 638.25334103 × 664 nm +
5506.80104833 × 689 nm − 1248.88851626 × 690 nm − 115.82768303 × 708 nm –
1517.88015709 × 709 nm − 2718.51566921 × 718 nm + 400.99891055 × 727 nm +
1629.99158725 × 728 nm + 1604.20664865 × 729 nm − 1048.07981365 × 730 nm +
396.39340421 × 757 nm − 433.694892 × 758 nm + 3035.53288044 × 759 nm –
614.25777917 × 762 nm − 1551.48090776 × 802 nm − 489.15224109 × 803 nm –
1372.19196525 × 804 nm − 2225.00027627 × 866 nm + 3936.97208447 × 867 nm +
61.51888687 × 868 n − 514.77643709 × 1178 nm − 1017.23089936 × 1179 nm +
1490.52821841 × 1232 nm + 422.56714205 × 1483 nm + 1043.34099139 × 1484 nm +
1378.72044837 × 1485 nm − 1762.26783483 × 1486 nm − 110.38296651 × 1487 nm –
1606.86292061 × 1488 nm + 420.02860207 × 1489 nm − 1060.13876294 × 1490 nm +
720.27928825 × 1491 nm − 28.76362342 × 1492 nm + 103.28184426 × 1493 nm –
131.95597966 × 1494 nm − 365.83963852 × 1793 nm − 553.20867642 × 1886 nm +
759.70141479 × 1887 nm + 1193.69613109 × 1888 nm − 3810.90239696 × 1898 nm +
1371.82998738 × 1899 nm + 2366.01957159 × 1907 nm − 1385.13255379 × 1908 nm +
1389.98364401 × 1911 nm + 1686.25002952 × 1912 nm − 590.34003403 × 1914 nm –
2568.58189163 × 1916 nm + 1891.41352763 × 1925 nm − 3160.17813263 × 1940 nm +
1545.65998348 × 1941 nm + 1350.36874433 × 1945 nm − 1110.56764511 × 1956 nm +
466.15005224 × 1989 nm − 263.86092155 × 1993 nm − 1999.25991196 × 2068 nm +
682.72535112 × 2069 nm − 1300.80069545 × 2070 nm + 2035.22693592 × 2071 nm –
533.64076109 × 2083 nm + 1647.95747007 × 2084 nm − 1711.27378902 × 2192 nm +
1915.97472243 × 2193 nm − 460.02914387 × 2360 nm − 1850.76561572 × 2364 nm –
2245.58661469 × 2365 nm + 3347.65479354 × 2368 nm + 648.05853418 × 2382 nm +
15.93562167 × 2385 nm + 34.51065143 × 2386 nm − 158.51284072 × 2389 nm +
1947.3063904 × 2390 nm − 1145.3684943 × 2394 nm − 3335.29857385 × 2426 nm +
3554.50781574 × 2427 nm − 2277.58894701 × 2428 nm + 3403.1296183 × 2429 nm +
1109.2586938 × 2430 nm − 3084.81572109 × 2431 nm + 1323.24564451 × 2432 nm +
3395.88193144 × 2433 nm − 5697.99831042 × 2434 nm + 3244.05960391 × 2450 nm –
695.62046814 × 2469 nm − 1721.5855356 × 2470 nm
pH Massey data = 5.63010896 + 1.01794868 × 359 nm + 11.21144696 × 360 nm –
7.32275159 × 361 nm − 21.40383301 × 389 nm − 14.03322979 × 390 nm + 30.74610746
× 391 nm − 43.53523382 × 392 nm + 18.23671329 × 394 nm − 6.38500327 × 395 nm +
18.01733161 × 396 nm − 27.36353051 × 432 nm − 19.20056161 × 439 nm + 92.96690364
× 461 nm + 48.16287377 × 516 nm − 43.75929762 × 517 nm − 45.84482625 × 518 nm –
16.79685965 × 522 nm − 6.78723094 × 536 nm − 25.46011315 × 537 nm + 68.40619006
× 538 nm − 24.08077525 × 576 nm − 16.97404877 × 577 nm + 8.06429782 × 578 nm –
67.30041419 × 579 nm − 6.13309954 × 580 nm + 38.00080165 × 581 nm + 19.49066231
× 596 nm + 57.02874678 × 597 nm + 56.66581798 × 598 nm + 40.13382241 × 608 nm +
11.51929348 × 609 nm − 5.23723548 × 610 nm − 115.34856771 × 611 nm + 27.83241776
× 634 nm − 21.40988878 × 685 nm − 66.58131083 × 686 nm + 37.08261582 × 701 nm +
20.93483536 × 702 nm + 25.60900836 × 703 nm − 16.63755709 × 704 nm − 7.58450299
× 731 nm − 73.19155201 732 nm + 3.76616034 × 733 nm + 9.37274134 × 767 nm +
56.97927421 × 768 nm + 43.02161764 × 769 nm − 10.15430763 × 784 nm − 34.43326738
× 785 nm − 26.49887985 × 786 nm + 129.32222188 × 891 nm + 3.23990609 × 929 nm –
68.25542179 × 930 nm − 45.04527251 × 931 nm + 13.84837643 × 1100 nm − 7.99616613
× 1101 nm + 51.69035257 × 1109 nm − 7.9522317 × 1110 nm − 58.35765755 × 1111 nm
+ 64.85300645 × 1115 nm + 30.3465026 × 1149 nm − 105.98328731 × 1170 nm –
31.56912981 × 1248 nm + 53.76128383 × 1251 nm + 21.86762012 × 1255 nm –
41.91309234 × 1259 nm − 22.43674708 × 1264 nm + 57.78670883 × 1265 nm –
67.27240656 × 1273 nm + 34.20628164 × 1274 nm − 45.64447098 × 1276 nm +
9.64816955 × 1277 nm − 3.06623343 × 1278 nm + 29.37988671 × 1280 nm +
36.03037732 × 1281 nm − 71.25645933 × 1284 nm + 6.79885619 × 1285 nm +
29.11077935 × 1288 nm + 42.77467661 × 1293 nm − 23.89894933 × 1294 nm –
1.03098712 × 1297 nm − 38.81837219 × 1300 nm + 61.98138819 × 1301 nm +
42.9406813 × 1304 nm + 37.08906097 × 1305 nm − 18.55261338 × 1308 nm –
82.81463312 × 1312 nm − 11.28607026 × 1418 nm + 45.80234711 × 1421 nm –
17.00111204 × 1422 nm − 81.73345042 × 1424 nm + 36.98733349 × 1425 nm +
92.21986869 × 1426 nm − 57.31536538 × 1434 nm + 47.75592428 × 1641 nm –
110.00831886 × 1645 nm + 69.69764291 × 1649 nm − 4.20631941 × 1889 nm –
23.3992297 × 2209 nm + 58.41679711 × 2218 nm − 34.627777 × 2229 nm
Ravensdown Olsen Sol. P ug/mL = 30.23032375 − 17395.35198111 × 515 nm +
16969.34406014 × 516 nm + 3377.61981422 × 549 nm + 52052.72258803 × 550 nm –
56378.23039846 × 551 nm − 10883.92875306 × 611 nm − 2804.46156902 × 612 nm +
31769.66357798 × 613 nm − 17718.32239769 × 614 nm − 35540.91236295 × 626 nm +
19983.25638365 × 627 nm + 37138.05492859 × 628 nm − 42450.34402066 × 629 nm +
29197.99472892 × 630 nm − 46675.44488017 × 662 nm + 94780.04643109 × 663 nm –
64090.24301552 × 664 nm + 29987.66554762 × 689 nm − 16542.41570829 × 690 nm –
24524.92814083 × 708 nm + 16852.13726875 × 709 nm + 13915.91992942 × 718 nm –
261970.26369825 × 727 nm + 475840.57721582 × 728 nm − 391631.84031072 × 729 nm
+ 167683.42635072 × 730 nm + 34174.1301415 × 757 nm − 55664.05487287 × 758 nm +
45346.43952343 × 759 nm − 24529.90590256 × 762 nm − 56908.44102923 × 802 nm +
68173.50089467 × 803 nm − 10648.33817098 × 804 nm − 45812.18514267 × 866 nm +
44779.116385 × 867 nm + 148.20876873 × 868 nm + 21509.19608288 × 1178 nm –
23210.76230177 × 1179 nm + 1205.11444786 × 1232 nm − 21251.86693495 × 1483 nm +
124470.56089342 × 1484 nm − 105176.50856368 × 1485 nm − 185934.27474499 × 1486
nm + 138981.68049312 × 1487 nm + 246634.81183744 × 1488 nm − 328405.42442083 ×
1489 nm + 347007.55586028 × 1490 nm − 292201.5362165 ×1491 nm + 200837.6178813
× 1492 nm − 179815.78120518 × 1493 nm + 55726.26983619 × 1494 nm − 16.73830056
× 1793 nm − 191.46145396 × 1886 nm − 15495.92099449 × 1887 nm + 20352.7752412 ×
1888 m − 19778.9353901 × 1898 nm + 13569.86772063 × 1899 nm + 13684.31307497 ×
1907 nm − 4213.77774362 × 1908 nm − 50337.04369486 × 1911 nm + 47183.0215495 ×
1912 nm − 7816.0179793 × 1914 nm + 5392.85376083 × 1916 nm − 2052.78826421 ×
1925 nm + 17278.11062412 × 1940 nm − 27761.29324011 × 1941 nm + 5853.66779632 ×
1945 nm + 5254.85824641 × 1956 nm + 3364.2802088 × 1989 nm − 3917.88758909 ×
1993 nm − 56380.67020239 × 2068 nm + 60708.68293837 × 2069 nm − 7333.4871515 ×
2070 nm − 1562.02303576 × 2071 nm + 15196.53356113 × 2083 nm − 11715.24142948 ×
2084 nm + 713.06495212 × 2192 nm − 1071.47197067 × 2193 nm − 2614.91400997 ×
2360 nm + 11410.26614702 × 2364 nm − 7602.59475447 × 2365 nm + 329.78450793 ×
2368 nm − 697.88676336 × 2382 nm − 5002.8475417 × 2385 nm + 6260.27394336 ×
2386 nm − 6485.55173936 × 2389 nm + 3766.47869786 × 2390 nm − 534.76176827 ×
2394 nm − 651.09384695 × 2426 nm + 14459.57991864 × 2427 nm − 8263.76328712 ×
2428 nm − 18615.1875937 × 2429 nm + 20821.74215637 × 2430 nm − 22094.66916225 ×
2431 nm + 19958.35600318 × 2432 nm + 5688.84328148 × 2433 nm − 9439.06667547 ×
2434 nm − 337.40531342 × 2450 nm + 1567.16270499 × 2469 nm − 1289.79561673 ×
2470 nm
Ravensdown pH = 6.18674446 − 1.69869607 × 359 nm + 6.31484627 × 360 nm +
0.08986659 × 361 nm − 10.19718185 × 389 nm + 57.97841255 × 390 nm − 88.63455731
× 391 nm + 29.6510304 × 392 nm − 3.20305907 × 394 nm + 9.54930601 × 395 nm –
1.69283658 × 396 nm + 20.93524516 × 432 nm − 6.77887439 × 439 nm − 45.13080489 ×
461 nm − 173.58845438 × 516 nm + 266.99997247 × 517 nm − 385.88722915 × 518 nm
+ 287.1056356 × 522 nm − 254.31109064 × 536 nm + 647.91459662 × 537 nm –
349.10130263 × 538 nm − 43.08258451 × 576 nm + 362.11769117 × 577 nm –
766.63930905 × 578 nm + 185.07826147 × 579 nm + 665.67198247 × 580 nm –
376.97383855 × 581 nm − 214.35928716 × 596 nm − 227.35866389 × 597 nm +
541.52766144 × 598 nm + 175.84690916 × 608 nm + 113.54131404 × 609 nm +
211.919296 × 610 nm − 594.98148968 × 611 nm − 51.3244556 × 634 nm + 284.70391609
× 685 nm − 306.84444538 × 686 nm + 1212.50847529 × 701 nm − 1936.03773524 × 702
nm + 1314.09886101 × 703 nm − 559.33951346 × 704 nm + 41.3318793 × 731 nm +
85.77972942 × 732 nm − 6.39724394 × 733 nm + 2.57821684 × 767 nm − 1000.33835402
× 768 nm + 638.43201143 × 769 nm + 497.76415139 × 784 nm − 500.55432586 × 785
nm + 178.4954138 × 786 nm + 278.09233576 × 891 nm − 97.70339466 × 929 nm +
17.59072331 × 930 nm − 139.22254262 × 931 nm − 338.49094149 × 1100 nm +
335.06610171 × 1101 nm + 192.94065592 × 1109 nm − 8.92408321 × 1110 nm –
215.20450248 × 1111 nm + 84.9204057 × 1115 nm + 105.04473744 × 1149 nm –
255.40995135 × 1170 nm + 510.86754127 × 1248 nm − 493.98138434 × 1251 nm +
131.95488055 × 1255 nm − 410.70269865 × 1259 nm − 425.9423493 × 1264 nm +
703.33488668 × 1265 nm − 880.17037685 × 1273 nm + 1217.18178423 × 1274 nm +
517.45423938 × 1276 nm − 2228.60786868 × 1277 nm + 1658.6689775 × 1278 nm +
405.04266763 × 1280 nm − 751.47374848 × 1281 nm + 1270.6998438 × 1284 nm –
1081.53285854 × 1285 nm − 94.52729087 × 1288 nm + 1303.6367605 × 1293 nm –
1595.58687516 × 1294 nm + 303.8508039 × 1297 nm + 219.35483385 × 1300 nm +
230.92135563 × 1301 nm − 857.25218971 × 1304 nm + 537.69713748 × 1305 nm +
190.37610831 × 1308 nm − 271.94718359 × 1312 nm + 543.88762491 × 1418 nm –
1825.22308523 × 1421 nm + 1399.02178773 × 1422 nm + 329.80032585 × 1424 nm –
4541.30142072 × 1425 nm + 4671.41597459 × 1426 nm − 578.23285744 × 1434 nm +
161.18160343 × 1641 nm − 116.28735521 × 1645 nm − 44.75511183 × 1649 nm –
15.22826404 × 1889 nm + 40.63028595 × 2209 nm − 72.16827033 × 2218 nm +
42.96064164 × 2229 nm
LogOlsenP Ravensdown Data = 1.41421696 − 262.38296498 × 515 nm + 258.3927565
× 516 nm + 172.63797279 × 549 nm + 565.81515751 × 550 nm − 760.83869552 × 551
nm − 265.81557966 × 611 nm + 352.59398899 × 612 nm + 30.70039851 × 613 nm –
64.88346955 × 614 nm − 720.64113906 × 626 nm + 342.74442767 × 627 nm +
917.56280873 × 628 nm − 1258.58083917 × 629 nm + 838.638906 × 630 nm –
977.90498588 × 662 nm + 1791.20280789 × 663 nm − 1142.04728078 × 664 nm +
446.0679803 × 689 nm − 148.19112984 × 690 nm − 232.03507741 × 708 nm –
11.00376379 × 709 nm + 252.55706692 × 718 nm − 4167.75593136 × 727 nm +
7798.83249416 × 728 nm − 6592.26457503 × 729 nm + 2897.4217993 × 730 nm +
240.60451648 × 757 nm − 107.42720557 × 758 nm + 258.77876235 × 759 nm –
471.60005204 × 762 nm − 748.4786656 × 802 nm + 951.61763404 × 803 nm −
167.32254136 × 804 nm − 797.01853402 × 866 nm + 857.16628768 × 867 nm –
76.77805452 × 868 nm + 217.55905357 × 1178 nm − 246.85313641 × 1179 nm +
22.81698457 × 1232 nm + 105.3998638 × 1483 nm + 1005.05780092 × 1484 nm –
621.61522117 × 1485 nm − 4286.12068281 × 1486 nm + 4315.85298814 × 1487 nm +
1035.3165668 × 1488 nm − 2177.24935658 × 1489 nm + 1891.09900621 × 1490 nm –
1646.62280566 × 1491 nm + 2219.08246915 × 1492 nm − 2146.86066432 × 1493 nm +
319.98869202 × 1494 nm − 11.55010056 × 1793 nm − 187.35487266 × 1886 nm +
411.3407968 × 1887 nm − 155.68803623 × 1888 nm − 230.23052205 × 1898 nm +
154.42632413 × 1899 nm − 49.4719615 × 1907 nm + 258.06274043 × 1908 nm –
1130.31963003 × 1911 nm + 1195.36384891 × 1912 nm − 285.57576757 × 1914 nm +
33.27687837 × 1916 nm + 35.10204655 × 1925 nm + 6.68288051 × 1940 nm –
146.78208608 × 1941 nm + 66.67246198 × 1945 nm + 40.53099809 × 1956 nm +
3.85709384 × 1989 nm − 1.53557831 × 1993 nm − 750.56274736 × 2068 nm +
906.91294082 × 2069 nm − 453.9769159 × 2070 nm + 238.11151817 × 2071 nm +
355.80990021 × 2083 nm − 314.89494526 × 2084 nm + 32.22063511 × 2192 nm –
41.55074839 × 2193 nm − 63.68003103 × 2360 nm + 248.87248573 × 2364 nm –
191.41192263 × 2365 nm + 37.99319427 × 2368 nm + 6.5000641 × 2382 nm –
97.59834886 × 2385 nm + 122.06277083 × 2386 nm − 149.60606473 × 2389 nm +
99.81677335 × 2390 nm − 18.9263297 × 2394 nm − 75.95234572 × 2426 nm +
413.74670128 × 2427 nm − 305.23617967 × 2428 nm − 170.12695893 × 2429 nm +
224.38494028 × 2430 nm − 345.81770354 × 2431 nm + 328.5991858 × 2432 nm +
144.86200043 × 2433 nm − 197.93334943 × 2434 nm − 9.76486789 × 2450 nm +
28.00901871 × 2469 nm − 15.8530771 × 2470 nm
Log Olsen P Massey Data = 0.88013594 − 17.65266195 × 515 nm + 32.64831668 × 516
nm − 23.87045858 × 549 nm − 24.62521319 × 550 nm + 11.81183849 × 551 nm –
16.76072924 × 611 nm − 19.90858466 × 612 nm + 39.71002999 × 613 nm + 24.8627212
× 614 nm + 56.8155205 × 626 nm − 13.74100331 × 627 nm − 29.22353436 × 628 nm +
5.16801954 × 629 nm + 21.54317538 × 630 nm + 2.35543078 × 662 nm − 134.35829477
× 663 nm + 4.10338132 × 664 nm + 93.51692756 × 689 nm + 10.70528884 × 690 nm +
5.94483344 × 708 nm − 17.02253754 × 709 nm − 37.71109024 × 718 nm + 30.38784528
× 727 nm − 6.34950395 × 728 nm + 12.8813182 × 729 nm − 20.96765759 × 730 nm +
53.39606312 × 757 nm − 11.07370831 × 758 nm + 29.08119915 × 759 nm –
18.69580045 × 762 nm − 43.03481018 × 802 nm + 10.56616882 × 803 nm –
49.96158803 × 804 nm − 11.93210691 × 866 nm + 69.1725134 × 867 nm − 8.91672726 ×
868 nm − 24.96420098 × 1178 nm − 17.85061366 × 1179 nm + 38.48982714 × 1232 nm
+ 34.2186828 × 1483 nm + 1.71967816 × 1484 nm + 35.11784548 × 1485 nm –
45.39746462 × 1486 nm − 3.33519592 × 1487 nm − 21.37228564 ×1488 nm +
5.09598897 × 1489 nm − 5.29940748 × 1490 nm + 2.53762136 × 1491 nm + 3.17278458
× 1492 nm + 9.30919722 × 1493 nm − 31.51731844 × 1494 nm − 2.73099208 × 1793 nm
− 7.66321304 × 1886 nm + 31.02393364 × 1887 nm − 0.48538538 × 1888 nm –
69.04625392 × 1898 nm + 9.87413403 × 1899 nm + 35.69223335 × 1907 nm +
15.37528336 × 1908 nm + 46.16127723 × 1911 nm + 24.82627982 × 1912 nm –
9.04673896 × 1914 nm − 57.03018431 × 1916 nm + 8.94977846 × 1925 nm –
61.10786769 × 1940 nm − 1.22981847 × 1941 nm + 59.27526757 × 1945 nm –
13.94808084 × 1956 nm − 23.16149111 × 1989 nm + 24.13821412 × 1993 nm –
39.09155072 × 2068 nm + 7.04608705 × 2069 nm + 7.43490488 × 2070 nm +
55.21290205 × 2071 nm − 14.29909075 × 2083 nm − 4.95445671 × 2084 nm –
31.51479668 × 2192 nm + 39.08059 × 2193 nm + 11.90398454 × 2360 nm –
35.53245923 × 2364 nm − 33.02879699 × 2365 nm + 61.63470833 × 2368 nm +
10.47308603 × 2382 nm + 9.11044008 × 2385 nm + 12.24158109 × 2386 nm –
20.01093487 × 2389 nm + 42.05763612 × 2390 nm − 56.47962304 × 2394 nm –
56.49844962 × 2426 nm + 47.71415088 × 2427 nm − 17.90993997 × 2428 nm +
43.29705927 × 2429 nm + 33.88853302 × 2430 nm − 80.79551169 × 2431 nm +
39.38333687 × 2432 nm + 44.5568069 × 2433 nm − 104.62940795 × 2434 nm +
61.64147234 × 2450 nm − 8.92227817 × 2469 nm − 22.35364028 × 2470 nm

References

  1. Cushnahan, M.Z.; Wood, B.A.; Yule, I.J. Is big data driving a paradigm shift in precision agriculture? In Proceedings of the 7th Asian-Australasian Conference on Precision Agriculture, Hamilton, New Zealand, 16–18 October 2017; p. 6. [Google Scholar]
  2. Grafton, M.C.E.; Manning, M.J. Establishing a Risk Profile for New Zealand, Pastoral Farms. Agriculture 2017, 7, 81. [Google Scholar] [CrossRef]
  3. Grafton, M.C.E.; Willis, L.A.; McVeagh, P.J.; Yule, I.J. Measuring pasture mass and quality indices over time using remote and proximal sensors. In Proceedings of the 13th International Conference on Precision Agriculture (Unpaginated, Online), St. Louis, MO, USA, 31 July–4 August 2016; p. 10. [Google Scholar]
  4. Chok, S.E.; Grafton, M.C.E.; Yule, I.J.; White, M. Accuracy of differential rate application technology for aerial spreading of granular fertilizer within New Zealand. In Proceedings of the 13th International Conference on Precision Agriculture (Unpaginated, Online), St. Louis, MO, USA, 31 July–4 August 2016; p. 13. [Google Scholar]
  5. Morton, J.D.; Baird, D.B.; Manning, M.J. A soil sampling protocol to minimise the spatial variability in soil test values in New Zealand hill country. N. Z. J. Agric. Res. 2000, 43, 367–375. [Google Scholar] [CrossRef]
  6. Kaul, T.M.C.; Grafton, M.C.E.; Hedley, M.J.; Yule, I.J. Understanding soil phosphorus variability with depth for the improvement of current soil sampling methods. In Science and Policy: Nutrient Management Challenges for the Next Generation; Occasional Report No. 30; Fertilizer and Lime Research Centre, Massey University: Palmerston North, New Zealand, 2017; p. 11. ISSN 0112-9902. [Google Scholar]
  7. Kaul, T.M.C.; Grafton, M.C.E. Geostatistical Determination of Soil Noise and Soil Phosphorus Spatial Variability. Agriculture 2017, 7, 83. [Google Scholar] [CrossRef]
  8. Stenberg, B.; Viscarra Rossel, R.A.; Mouazen, A.M.; Wetterlind, J. Visible and Near Infrared Spectroscopy in Soil Science. In Advances in Agronomy, 1st ed.; Sparks, D.L., Ed.; Academic Press: Burlington, NJ, USA, 2010; pp. 163–215. [Google Scholar] [CrossRef]
  9. Viscarra Rossel, R.A.; Behrens, T. Using data mining to model and interpret soil diffuse reflectance spectra. Geoderma 2010, 158, 46–54. [Google Scholar] [CrossRef]
  10. Nocita, M.; Stevens, A.; Toth, G.; Panagos, P.; van Wesemael, B.; Montanarella, L. Prediction of soil organic carbon content by diffuse reflectance spectroscopy a local partial least square regression approach. Soil Biol. Biochem. 2014, 68, 337–347. [Google Scholar] [CrossRef]
  11. Roudier, P.; Hedley, C.B.; Ross, C.W. Prediction of Volumetric Soil Organic Carbon from Field Moist Intact Soil Cores. Eur. J. Soil Sci. 2015, 66, 651–660. [Google Scholar] [CrossRef]
  12. Wijewardane, N.K.; Ge, Y. Laboratory evaluation of two VNIR optical sensor designs for vertical soil sensing. In Proceedings of the 13th International Conference on Precision Agriculture (Unpaginated Online), St. Louis, MO, USA, 31 July–4 August 2016. [Google Scholar]
  13. Cushnahan, T.A.; Yule, I.J.; Grafton, M.C.E.; Pullanagari, R.; White, M. The Classification of Hill Country Vegetation from Hyperspectral Imagery. In Science and Policy: Nutrient Management Challenges for the Next Generation; Currie, Occasional Report No. 30; Currie, L.D., Hedley, M.J., Eds.; Fertilizer and Lime Research Centre, Massey University: Palmerston North, New Zealand, 2017; p. 10. ISSN 0112-9902. [Google Scholar]
  14. Murphy, J.; Riley, J.P. A modified single solution method for the determination of phosphate in natural waters. Anal. Chim. Acta 1962, 27, 31–36. [Google Scholar] [CrossRef]
  15. Minasny, B.; McBratney, A.B.; Bellon-Maurel, V.; Roger, J.-M.; Gobrecht, A.; Ferrand, L.; Joalland, S. Removing the effect of soil moisture from NIR diffuse reflectance spectra for the prediction of soil organic carbon. Geoderma 2011, 67, 118–124. [Google Scholar] [CrossRef]
  16. Team, R.C. R: A Language and Environment for Statistical Computing; R Foundation for Statistical Computing: Vienna, Austria, 2016. [Google Scholar]
  17. Mevik, B.H.; Wehrens, R. The pls package: Principal component and partial least squares regression in R. J. Stat. Soft. 2007, 18, 24. Available online: https://repository.ubn.ru.nl/bitstream/handle/2066/36604/36604.pdf (accessed on 20 January 2019). [CrossRef]
  18. Morton, J.D.; Sinclair, A.G.; Morrison, J.D.; Smith, L.C.; Dodds, K.G. Balanced and adequate nutrition of phosphorus and sulphur in pasture. N. Z. Agric. Res. 1998, 41, 487–496. [Google Scholar] [CrossRef]
  19. Cornforth, I.S.; Sinclair, A.G. Fertiliser and Lime Recommendations for Pastures and Crops in New Zealand; Ministry of Agriculture and Fisheries: Wellington, New Zealand, 1984; p. 76.
  20. Lucci, G.M.; McDowell, R.W.; Condron, L.M. Potential phosphorus and sediment loads from sources within a dairy farmed catchment. Soil Use Manag. 2010, 26, 44–52. [Google Scholar] [CrossRef]
Figure 1. Shows, the location and 30-year, average annual rainfall for NZ (1981–2010). Source: National Institute of Water and Atmosphere NZ. https://www.niwa.co.nz/climate/national-and-regional-climate-maps/national accessed 7 March 2019.
Figure 1. Shows, the location and 30-year, average annual rainfall for NZ (1981–2010). Source: National Institute of Water and Atmosphere NZ. https://www.niwa.co.nz/climate/national-and-regional-climate-maps/national accessed 7 March 2019.
Agriculture 09 00055 g001
Figure 2. Location and name of farms where samples were collected. The sampling body (Ravensdown Ltd, or Massey University) the size of the farm and number of samples taken.
Figure 2. Location and name of farms where samples were collected. The sampling body (Ravensdown Ltd, or Massey University) the size of the farm and number of samples taken.
Agriculture 09 00055 g002
Figure 3. (a) Shows the distribution of the Massey Olsen P data; (b) shows the distribution of Massey pH data; (c) shows the distribution of Ravensdown Olsen P data; and (d) shows the distribution of Ravensdown pH data.
Figure 3. (a) Shows the distribution of the Massey Olsen P data; (b) shows the distribution of Massey pH data; (c) shows the distribution of Ravensdown Olsen P data; and (d) shows the distribution of Ravensdown pH data.
Agriculture 09 00055 g003aAgriculture 09 00055 g003b
Figure 4. (a) shows a Q-Q Normality test for the raw Olsen P data with a Chi2 distribution. (b) Shows that the Log10 of the Olsen P of the data produces a normal distribution.
Figure 4. (a) shows a Q-Q Normality test for the raw Olsen P data with a Chi2 distribution. (b) Shows that the Log10 of the Olsen P of the data produces a normal distribution.
Agriculture 09 00055 g004
Figure 5. Scatter plots of Olsen P and pH modelled results compared to measured, for (a,b) Massey calibration set and (cf) application of the Massey model to the Ravensdown set of data for validation and comparison to the data set regression.
Figure 5. Scatter plots of Olsen P and pH modelled results compared to measured, for (a,b) Massey calibration set and (cf) application of the Massey model to the Ravensdown set of data for validation and comparison to the data set regression.
Agriculture 09 00055 g005aAgriculture 09 00055 g005bAgriculture 09 00055 g005c
Table 1. Mean and standard deviations of the two data sets.
Table 1. Mean and standard deviations of the two data sets.
Data SetMeanStandard Deviation
Massey Olsen P18.113.9
Massey pH5.550.33
Ravensdown Olsen P23.115.4
Ravensdown pH5.720.32
Table 2. Summary of regression analyses.
Table 2. Summary of regression analyses.
Data SetAdj. R2Std. ErrFp Value
Ravensdown Olsen P 10.4711.212.319<2.2 × 10−16
Ravensdown pH 10.710.174.688<2.2 × 10−16
Massey pH 20.420.257.3<2.2 × 10−16
Massey pH 30.420.257.3<0.0001
Massey Olsen P 30.569.0311.99<0.0001
Ravensdown pH 40.420.2424.53<0.0001
Ravensdown Olsen P 40.2713.2112.53<0.0001
Ravensdown Log OP 40.349.0317.42<0.0001
Massey Log Olsen P 40.460.258.5<0.0001
1 R regression 2148 and 1040 degrees of freedom; 2 R regression 100 and 782 degrees of freedom; 3 StatTools regression 100 and 782 degrees of freedom; and 4 StatTools regression 100 and 3088 degrees of freedom.

Share and Cite

MDPI and ACS Style

Grafton, M.; Kaul, T.; Palmer, A.; Bishop, P.; White, M. Technical Note: Regression Analysis of Proximal Hyperspectral Data to Predict Soil pH and Olsen P. Agriculture 2019, 9, 55. https://0-doi-org.brum.beds.ac.uk/10.3390/agriculture9030055

AMA Style

Grafton M, Kaul T, Palmer A, Bishop P, White M. Technical Note: Regression Analysis of Proximal Hyperspectral Data to Predict Soil pH and Olsen P. Agriculture. 2019; 9(3):55. https://0-doi-org.brum.beds.ac.uk/10.3390/agriculture9030055

Chicago/Turabian Style

Grafton, Miles, Therese Kaul, Alan Palmer, Peter Bishop, and Michael White. 2019. "Technical Note: Regression Analysis of Proximal Hyperspectral Data to Predict Soil pH and Olsen P" Agriculture 9, no. 3: 55. https://0-doi-org.brum.beds.ac.uk/10.3390/agriculture9030055

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop