Spatial fuzzy c-means for delineation of potential management zones based on soil physico-chemical properties and cowpea yield

doi:10.21203/rs.3.rs-1852204/v1

Download PDF

Research Article

Spatial fuzzy c-means for delineation of potential management zones based on soil physico-chemical properties and cowpea yield

https://doi.org/10.21203/rs.3.rs-1852204/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Aims Management zone delineation is one of the principles of precision agriculture. The spatial variability of the different environmental factors (soil and plants) is evident in several studies. The associations between soil physico-chemical properties and cowpea yield have been assessed, but scholars have usually reported a large unexplained variance of ranges between kriged maps. This suggests that a deeper exploration into the soil properties of these spatial interactions will help develop our understanding on how to reduce the number of soil property maps to delineate management zones and simplify interpretation.

Methods This research proposes to investigate whether factorial kriging analysis can be used as an auxiliary variable in cokriging to estimate soil properties and cowpea, and what is the potential of spatial Fuzzy C-Means associated with factorial kriging analysis to delineate management zones.

Results By using factor maps and spatial clustering, two management zones were classified. The kriged maps of soil properties were not efficient to delineate management zones.

Conclusions Appropriately the results indicate that the proposed method is a practical tool to delineate management zones and it is more efficient than multiple soil properties maps.

Vigna unguiculata

Geostatistics

Factorial kriging analysis

Precision agriculture

Spatial variability

Table 1

Criteria for defining management zones based on a cowpea yield map according to Milani et al. (2006).

Management zones	Management requirements	Zones
Low yield zone	< 95% in relation for medium	Low
Medium yield zone	95% Medium yield 105%	Medium
High yield zone	< 105% in relation for medium	High

Table 2

Descriptive statistics of soil properties and cowpea yield (n=66).

Properties	Min	Max	Mean	SD	CV	Sk	Kurtosis	p-value
Cy	10.76	110.19	64.97	20.36	31.34	-0.20	0.42		>0.100*
P	0.01	0.09	0.04	0.02	38.59	0.56	-0.50		0.040
K	0.00	0.17	0.07	0.04	55.33	0.29	-0.01		>0.100*
Na	0.00	0.02	0.01	0.01	45.57	-0.56	0.04		>0.100*
Al	0.38	1.19	0.87	0.18	21.22	-0.62	0.11		0.098*
Ca	0.08	1.15	0.35	0.22	62.50	1.68	2.92		<0.010
Mg	0.07	0.60	0.19	0.10	51.96	1.51	3.58		<0.010
Sand	505.00	892.00	835.29 49.56	48.62	5.82	-5.00	33.11		<0.010
Silt	8.00	171.00	49.56	21.39	43.16	2.90	15.50		<0.010
Clay	60.00	160.00	110.00	22.80	20.73	0.19	-0.21		>0.100*

Cy: Cowpea yield, SD: standard deviation, CV: coefficient of variation, Sk: Skewness, p-value: *Normal distribution by the Shapiro-Wilk test when detected p-value >0.05.

Table 3

Summarization of factorial analysis and correlations between soil characteristics and rotated factors (F).

	Factor 1	Factor 2	Factor 3	Factor 4
Eigenvalues	2.84	1.71	1.45	1.15
Cumulative variance (%)	31.61	50.66	66.83	79.71
Explained variance (%)	31.61	19.05	16.17	12.88
Correlations
P	-0.19	-0.07	0.58	-0.12
K	-0.50	0.72	-0.27	-0.08
Na	-0.12	0.90	-0.13	0.10
Al	0.85	0.28	-0.01	0.03
Ca	-0.78	-0.42	-0.15	-0.23
Mg	-0.81	0.08	-0.20	-0.42
Sand	-0.44	0.16	0.66	0.31
Silt	-0.07	-0.29	-0.71	0.49
Clay	0.57	0.03	-0.12	-0.75
	Al + Ca + Mg	K + Na	Silt	Clay
	Chemical properties		Soil texture

Correlations with absolute values higher than 0.7 are in bold.

Table 4

Spatial diagnostic for the optimal number of clusters (MZs) with the explained inertia, silhouette index and Spatial inconsistency applied in RGB maps of Factors.

	FCM	GFCM	sFCM	Factors	Numbers of MZs
Explained inertia	0.38	0.42	0.39	F123
Silhouette index	0.65	0.63	0.65		2
Spatial inconsistency	0.08	0.15	0.08
Explained inertia	0.27	0.28	0.28	F124
Silhouette index	0.59	0.59	0.59		2
Spatial inconsistency	0.06	0.06	0.06

FCM: Fuzzy c-means; GFCM: Generalized fuzzy c-means; sFCM: Spatial

Graphicalabstract01.tif

Download PDF

Version 1

posted

You are reading this latest preprint version

Spatial fuzzy c-means for delineation of potential management zones based on soil physico-chemical properties and cowpea yield

Status:

Version 1

Abstract

Figures

Full Text

Tables

Supplementary Files

Status:

Version 1