Changes in sorghum production, soil P forms and P use efficiency following long-term application of manure, compost and straw in a Ferric Lixisol

ABSTRACT

Low phosphorus availability in cultivated soils limits sustainable crop production in sub‐Saharan Africa. This study aimed at evaluating the effect of long‐term application of different types of organic amendments on soil P forms, P use efficiency and sorghum yields. A long term experiment established in 1980 at Saria in Burkina Faso, comparing the effects of manure, compost and sorghum straw was used. Manure and compost significantly increased organic P and resin‐P by about 35% and 64%, respectively after 10 and 32 years of sorghum cultivation, and HCl‐P after 32 years of cultivation compared to the control. Manure significantly increased NaHCO₃-Pi and NaOH-Pi by 63% and 26%, respectively compared to the control. Sorghum straw had little effect on measured soil P forms. Manure and compost were the best in increasing sorghum grain yield, which effect were strongly correlated to soil pH, carbon and nitrogen. The partial factors productivities of P resulting from the application of studied organic amendments were similar and low, but significantly higher than that of the control treatment. Organic amendments with high P content, maintaining soil carbon and pH could be used to improve soil P availability, sorghum yield and reduce the demand for mineral phosphorus fertilizers.

KEYWORDS:

• Organic amendments
• phosphorus fractions
• phosphorus use efficiency
• sorghum production
• Lixisol

Introduction

Soil phosphorus replenishment is a plausible entry point to achieve the green revolution in sub-Saharan Africa (Mokwunye and Bationo, 2011). Phosphorus is indeed one of the main factors that limit crop production in most of cultivated soils of sub-Saharan Africa (Sedogo 1993; Bationo et al. 1998). In these soils, total and available phosphorus are less than the required amounts for optimum plant growth and yield (Compaoré et al. 2003; Lompo 2009). Therefore, the need to improve soil P availability is seen as crucial to increase and sustain soil and crop productivity. Water soluble P fertilisers could be used to improve soil phosphorus status. However, applying mineral fertilisers solely on lixisols led to a degradation of their properties and decreased crop yields after 5 years of application (Sedogo 1993). Furthermore, the primary source of water soluble P fertilisers which is phosphate rock is a non-renewable resource and a possible phosphorus shortage in next centuries is actually on the centre of debates (Van Kauwenbergh 2006; Cordell et al. 2009; Van Kauwenbergh et al. 2013). There is therefore a dire need to develop and promote soil phosphorus management practices which increase the effectiveness of mineral phosphorus fertilisers while reducing their demand.

The integrated use of mineral fertilisers and organic amendments has been shown as an effective way to improve soil P availability and crops yields. Organic amendments can indeed increase soil P availability by adding P in soil after mineralisation or by chelating P fixating compounds like iron, calcium and aluminum (Marschner and Rengel 2012). Whereas organic amendments are quite important in increasing soil P availability, their quality strongly determine their effects on soil properties, plant P uptake and crop yields. It has been shown that P release in soil from organic residues is linked to residues’ P, lignin and polyphenol content and C/P ratio (Kwabiah et al. 2003; Baggie et al. 2004; Nwoke et al. 2004). Although the release of P from various residues is well documented, few studies have been conducted on the long term effect of those various organic amendments on acid tropical soils P fractions and P use efficiency. Long-term experiments are good basis for understanding processes that lead to soil fertility changes and allow the identification of sustainable management practices of soil fertility (Richter et al. 2007; Bationo et al. 2012). There is also evidence from research that investigating on the dynamics of soil P fractions is a useful approach for determining the fate of applied P in agricultural soils (Negassa and Leinweber 2009). The objectives of this study were therefore to (1) determine the changes of selected soil phosphorus pools as affected by the long-term application of different types of organic amendments and (2) investigate the effect of long-term application of these types of organic amendments on sorghum yield, P uptake and P use efficiency.

Materials and methods

Experimental site

The long-term experiment is implemented at the research station of ‘Institut de l’environnement et de Recherches Agricoles’ (INERA) at Saria in the Central West area of Burkina Faso (Latitude 12° 16′ N, Longitude 2° 9′ W, Altitude 300 m) on a lixisol. Saria is located within the North Sudanian Zone of Burkina Faso. The annual mean rainfall is 800 mm fluctuating between 600 and 1200 mm over the years (Figure 1). The average minimum temperature in Saria is about 18°C while the maximum is about 45°C. The average potential evapotranspiration is about 2000 mm. The soil types of the experimental site are Lixisols which represent about 39% of soils in Burkina Faso. Those soils are sandy and have a very fragile structure, very low organic matter and available P contents (Sedogo 1993; Compaoré et al. 2003).

Figure 1. Trend of sorghum grain yields as affected by long term application of different types of organic amendments from 1980 to 2012 in Saria, Burkina Faso.

Experimental design and management

The long-term field trial of Saria was implemented in 1980 to compare the effect of different types of organic amendments on soil properties and sorghum yield. The experiment was originally laid out in a randomised block design with two factors (the type of organic amendment, the mineral fertilisation) and six blocks. This study focused on four treatments of organic amendments which were: Control (no organic amendment), Sorghum straw, Manure, and Compost, each in combination with one treatment of mineral fertilisation (60 kg N ha⁻¹ + 30 kg P₂O₅ ha⁻¹ + 30 kg K₂O ha⁻¹). The compost is produced by composting during three months a mixture of 75% sorghum straw and 25% manure. The chemical characteristics of the applied organic amendments in 2012 are shown in the Table 1. The experimental plot size is 20.8 m² (5.2 m × 4 m). Since 1980, at the beginning of each rainy season, prior to sowing, organic amendments are applied at the rate of 10 t dry matter ha⁻¹, and plots are tilled down to 20 cm mechanically. Then, sorghum (Sorghum Bicolor (L.) Moench) is sown at a plant spacing of 80 cm × 40 cm. From the beginning of the trial to 2012, sorghum varieties used in this experiment differed by their yield potential and by their sensitivity to photoperiod. (For example from 1980 to 1987 the sorghum variety was E35-1, a photoperiod sensitive variety, in 1988, 1989, 1990 and 2012 the sorghum variety was ICSV 1049, a photoperiod insensitive variety). Phosphorus and potassium mineral fertilisers are applied as Triple Super Phosphate (TSP, 46% P₂O₅) and potassium chloride (KCl, 60% K₂O) respectively 2 weeks after sowing by broadcasting. Nitrogen mineral fertiliser is applied as urea (46% N) by broadcasting in two equal fractions, the first fraction being applied 2 weeks after sowing, the second fraction being applied at the boot stage of sorghum. The weeding is done when necessary.

Table 1. Chemical characteristics of the applied sorghum straw, manure and compost in the long term experiment in 2012 in Saria, Burkina Faso.

	pH	C	N	P	K	C/N	C/P
Organic amendments		g kg dry matter^−1
Sorghum straw	6.1	425	2.3	0.23	2.79	184.8	1847.8
Farmyard manure	7.7	151	13.4	2.34	8.61	11.3	64.8
Compost	7.5	164	9.7	1.54	8.08	16.9	106.5

Sampling and analyses of soil, plant and organic amendments

We worked with soil samples from 2012 and 1990. In 2012, soil of each experimental plot was sampled at harvest from five spots at a depth of 0–20 cm. Samples collected from the five spots were mixed to obtain composite samples. Soils samples were then air-dried at room temperature and sieved through a 2 mm mesh and stored in plastic bags for future chemical analysis. The soil samples of the year 1990 were collected from the soil bank of the laboratory of INERA, which soils were previously collected and stored as stated above.

Soil analyses were done on three blocks of the six total blocks of the experimental trial. Soil pH (H₂O) was measured with an electronic pH-meter (WTW InoLab, Weilheim, Germany) in slurries formed from a ratio of one gram of soil to 2.5 ml of distilled water according to Afnor (1981). The soil total carbon was determined using the protocol of Walkley and Black (1934). Soil total nitrogen and total phosphorus were measured using an automatic colorimeter SKALAR (Skalar SANplus Segmented flow analyzer, Model 4000-02, Breda, Holland) after digestion of one gram of soil sample with concentrated sulfuric acid in the presence of hydrogen peroxide, selenium and salicylic acid (Okalebo et al. 2002). Soil organic phosphorus was determined by the method of Saunders and Williams (Saunders and Williams 1955) modified by Walker and Adams (1958).

The procedure of Hedley et al. (1982) as modified by Condron and Goh (1989) was used to sequentially determine soil inorganic P fractions. Four grams of soil was sequentially extracted in the following order: anion exchange resin in the bicarbonate form; sodium bicarbonate (0.5 M NaHCO_3, pH 8.5); sodium hydroxide solution (0.1 M NaOH and hydrochloride acid (1M HCl). Phosphorus content of the extract was measured using a manual colorimeter (Cecil 3021) at 880 nm following the Murphy and Riley method (Murphy and Riley 1962).

For the determination of P contents of sorghum straw and grain, a quantity of 200–500 mg of material was placed in an electric furnace at 550°C for 5 h. Phosphorus in the ash materials was extracted with 2 ml of concentrated (65%) HNO₃ and the volume made up to 100 ml with distilled water. Phosphorus concentration in the extract was measured on a manual colorimeter (Cecil 3021) at 880 nm following the Murphy and Riley method (Murphy and Riley 1962).

The total organic matter in the different types of organic amendments was determined from mass loss after ignition at 550°C in the electric furnace. The total nitrogen and phosphorus of organic amendments were determined as was done for the soil samples. The total potassium in organic amendments was determined using a flame photometer (Jencons PFP 7, Jenway LTD, Felsted, England) after the same procedure of digestion used in the determination of total nitrogen and phosphorus. The pH of each organic amendment was also determined as done for soil but from a ratio of 1 g substrate to 5 ml water.

Statistical analysis and calculations

The data were analyzed using the GENSTAT software edition 9. Two-way ANOVA (in randomised blocks) was used to analyze soil data considering the type of the organic amendment and the year of soil sampling as the two factors. One-way ANOVA (in randomised block) was used for yield data and P uptake data. The least significant difference (Lsd) was used to compare the means at 95% confidence level when the factor effect was significant. The data of partial factor productivity of P was log transformed for variance homogeneity before variance analysis. A principal component analysis was done with the software XLSTAT 2007 to classify the treatments according to the measured parameters and to identify the parameters that determine sorghum yield increases and P use efficiency.

P use efficiency was evaluated by determining the partial agronomic factor of applied P according to Syers et al. (2008) as follows:The total P applied is the sum of the P applied with the mineral fertiliser and the organic amendment.

The P partial balance in 2012 was calculated as follows:

Results

Effect of organic amendments on soil chemical properties

There was a significant interaction between the application of organic amendment and the year of soil sampling for soil pH and organic carbon (p ≥ 0.05) (Table 2). Soil pH was generally acidic in all treatments and in both years and varied from 5.87 to 6.29 in 1990 and from 4.55 to 5.85 in 2012. In 1990, all the three types of organic amendment treatments recorded similar soil pH which was significantly higher than the control. In 2012, the soil pH ranked as follows: control < straw < compost = manure. The increase over the control in soil pH represented 0.77, 1.29, 1.25 units for treatments that received straw, compost and manure, respectively. From the year 1990 to year 2012, soil pH decreased in all treatments from 1.32, 0.96, 0.46, 0.38 units for the control, the straw, compost, and manure treatments, respectively.

Table 2. Effects of different types of organic amendments and age of the long term experiment on selected soil chemical characteristics in Saria, Burkina Faso.

Treatments	pH	C	N
		(g kg^−1)
Organic amendments (OA)
Control	5.21	2.27	0.20
Compost	6.03	3.78	0.37
Manure	6.04	4.26	0.42
Straw	5.81	3.51	0.32
P OA	<0.001	<0.001	<0.001
LSD OA	0.16	0.56	0.06
Year of soil sampling
1990	6.16	4.07	0.41
2012	5.38	2.84	0.25
P Age	<0.001	<0.001	<0.001
LSD Age	0.12	0.39	0.04
OA × Age
Control × 1990	5.87	2.92	0.27
Control × 2012	4.55	1.62	0.14
Compost × 1990	6.26	4.06	0.44
Compost × 2012	5.80	3.49	0.31
Manure × 1990	6.22	4.66	0.51
Manure × 2012	5.85	3.85	0.33
Straw × 1990	6.29	4.62	0.43
Straw × 2012	5.33	2.4	0.21
P OA × Age	<0.001	0.032	0.333
LSD OA × Age	0.23	0.79	NS
CV (%)	2.3	13.0	14.6

NS: Not Significant.

Soil organic carbon ranged from 2.92 g kg⁻¹ to 4.66 g kg⁻¹ in 1990 and from 1.62 g kg⁻¹ to 3.85 g kg⁻¹ in 2012. For both years, the treatments receiving organic amendments recorded higher soil organic matter content compared to the control. In 1990, there were no significant differences in soil organic carbon content among the different types of organic amendments. In 2012, soil organic carbon content ranked as follows: control < straw < compost = manure. Treatments receiving sorghum straw, composts and manure resulted to 48%, 115%, 138%, respectively more soil organic carbon than the control. The decrease in soil organic carbon content from 1990 to 2012 amounted to 14%, 17%, 48%, and 45% for the compost, manure, control, and straw treatments, respectively but this decrease was not significant for compost and manure treatments.

There was not a significant interaction between the effect of the different types of organic amendment and the year of soil sampling for soil total nitrogen content. The soil nitrogen contents varied from 0.27 g kg⁻¹ to 0.51 g kg⁻¹ in 1990 and from 0.14 g kg⁻¹ to 0.33 g kg⁻¹ in 2012 (Table 2). In both years, all organic amendment treated plots recorded higher soil total N than the control treatment with manure as the best organic amendment followed by compost. Soil total nitrogen decreased by about 39% from 1990 to 2012 for all treatments.

Effect of organic amendments on soil phosphorus forms

The Table 3 presents the effect of different type of organic amendments on soil P forms and fractions. There was no significant interaction between the effects of the application of the organic amendments and the year of soil sampling for soil total P, organic P, Resin-P, NaHCO₃-Pi and NaOH-Pi (p > 0.05).

Table 3. Effect of long term application of different types of organic amendments on soil P fractions in Saria, Burkina Faso.

Treatments	Resin-P	NaHCO3-Pi	NaOH-Pi	HCl-P	Organic P	Total P
	mg kg^−1
Organic amendments (OA)
Control	4.92	6.07	19.87	5.08	40.95	93.6
Compost	7.58	7.98	21.99	9.73	53.92	161.7
Manure	8.58	9.88	24.99	10.54	56.63	160.7
Sorghum Straw	5.16	5.70	18.67	4.15	46.85	119.3
P OA	0.009	0.005	0.007	<0.001	<0.001	<0.001
LSD OA	2.275	2.29	3.424	2.147	5.097	23.28
Year of soil sampling
1990	4.69	5.56	16.89	5.35	49.03	101.7
2012	8.43	9.25	25.86	9.39	50.14	169.0
P Age	<0.001	<0.001	<0.001	<0.001	0.520	<0.001
LSD Age	1.609	1.619	2.421	1.518	NS	16.46
OA × Age
Control × 1990	3.78	3.73	16.83	6.8	43.10	69.5
Control × 2012	6.05	8.4	22.9	3.35	38.80	117.7
Compost × 1990	4.73	6	16.95	6.7	52.07	114.2
Compost × 2012	10.43	9.95	27.03	12.75	55.77	209.2
Manure × 1990	6.3	7.53	19.33	4.53	54.40	114.4
Manure × 2012	10.85	12.23	30.64	16.54	58.87	207.0
Sorghum Straw × 1990	3.93	4.98	14.45	3.38	46.57	96.4
Sorghum Straw × 2012	6.38	6.41	22.88	4.91	47.13	142.2
P OA × Age	0.336	0.404	0.419	<0.001	0.283	0.061
LSD OA × Age	NS	NS	NS	3.037	NS	32.93
CV (%)	28	25	12.9	23.5	8.3	14.1

NS: Not Significant.

The soil total P ranged from 69.5 mg kg⁻¹ in the control treatment to 114 mg kg⁻¹ in the manure treatment in 1990 and from 118 mg kg⁻¹ in the control treatment to 209 mg kg⁻¹ in the compost treatment in 2012. Manure and compost significantly increased soil total P by about 62% in average relative to that of the control treatment. However, differences between total soil P due to application of sorghum straw and control were not significant. From 1990 to 2012, significant (p < 0.05) increases in soil total P averaged 66% for all treatments.

Soil organic P varied from 43.1 mg kg⁻¹ in the control treatment to 54.4 mg kg⁻¹ in the manure treatment in 1990 and from 38.8 mg kg⁻¹ in the control treatment to 58.8 mg kg⁻¹ in the manure treatment in 2012. The organic amendments significantly increased soil organic P compared to the control (p < 0.05) and ranged as follow: Control < Sorghum Straw < Compost = Manure. The increase above that of the control was about 35% for manure and compost and 14% for sorghum straw. There was not a significant difference between soil organic P measured in 1990 and those measured in 2012.

The Resin-P fraction ranged from 3.78 mg kg⁻¹ in the control treatment to 6.3 mg kg⁻¹ in the manure treatment in 1990 and from 6.05 mg kg⁻¹ in the control treatment to 10.8 mg kg⁻¹ in the manure treatment in 2012. The compost and manure application increased significantly soil resin-P concentration by 64% in average over that of the control treatment, which received only mineral fertiliser. The application of sorghum straw did not significantly change soil resin-P compared to that of the control treatment. From 1990 to 2012 soil Resin P was significantly (p < 0.05) increased by about 80% for all treatments.

The NaHCO₃-Pi fraction varied from 3.73 mg kg⁻¹ in the control treatment to 7.53 mg kg⁻¹ in the manure treatment in 1990 and from 8.4 mg kg⁻¹ in the control treatment to 12.2 mg kg⁻¹ in the manure treatment in 2012. Only the manure treatment significantly increased this fraction and by 63% over the control treatment. From 1990 to 2012 NaHCO₃-Pi was significantly (p < 0.05) increased by about 66% for all treatments.

The NaOH-Pi was the highest fraction of the studied inorganic phosphorus pools. Soil NaOH-Pi ranged from 14.5 mg kg⁻¹ in sorghum straw treated plots to 19.3 mg kg⁻¹ in manure treated plots in 1990 and from 22.8 mg kg⁻¹ in sorghum straw treated plots to 30.6 mg kg⁻¹ in manure treated plots in 2012. The application of manure significantly increased this fraction and by 26% over the control treatment. The change in the NaOH-Pi fraction resulting from the other types of organic amendments was not significantly different from that of the control treatment. From 1990 to 2012 NaOH-Pi significantly (p < 0.05) increased by about 53% for all treatments.

There was a significant interaction between the effects of the application of different types of organic amendments and the year of soil sampling for soil HCl-P (Table 3). HCl-P fraction varied from 3.38 mg kg⁻¹ in sorghum straw treated plots to 6.8 mg kg⁻¹ in the manure treated plots in 1990 and from 3.35 mg kg⁻¹ in the control treatment to 16.5 mg kg⁻¹ in the manure treated plots in 2012. In 1990, none of the three organic amendments significantly increased the proportion of HCl-P fraction compared to the control. On the contrary, the application of straw led to a decrease of HCl-P fraction compared to the control. In 2012, compost and manure amendments increased HCl-P fraction by 18% and 34%, respectively while the difference between the straw amended soil and the control was not significant at p < 0.05. From 1990 to 2012 HCl-P increased by 265% and 90%, respectively in the manure and compost treated plots while it decreased in the control treatment by 51% and did not significantly change in the sorghum straw treatment.

Effects of organic amendments on sorghum yield and phosphorus uptake and use efficiency and phosphorus partial balance

Figure 1 presents the trend of sorghum yield as affected by long-term application of the different types of organic amendments. The sorghum yields varied throughout the years with the highest yields being obtained from the manure and compost (about 3700 kg ha⁻¹) amended treatments while the lowest yields were obtained from the control treatment plots (26 kg ha⁻¹). The statistical analysis done by regrouping sorghum grain yield by time frame of about ten year (Table 4) shows that from 1982 to 1990 sorghum straw and manure were the best organic amendments for increasing sorghum grain yield. However, during the two following last decades (1991–2000 and 2001–2012), sorghum straw had little effect in increasing sorghum yield while manure and compost were the best organics amendments for increasing sorghum grain yields above the control treatment.

Table 4. Change in Sorghum grain yield response to the application of different type of organic amendments kg ha⁻¹ in Saria, Burkina Faso.

Treatments	Mean 1982–1990	Mean 1991–2000	Mean 2001–2012
	kg ha^−1
Control	1,653	1,526	864
Compost	2,001	2,345	2,345
Manure	2,162	2,444	2,002
Sorghum Straw	2,307	1,965	1,376
p Values	0.033	0.009	0.003
LSD (0.05)	434.4	535.9	556.4
CV (%)	17.4	21	29.8

The Table 5 presents the effect of different type of organic amendments on sorghum P uptake, partial factor productivity and P partial balance in 2012.

Table 5. Sorghum P uptake and use efficiency and P partial balance as affected by long term application of different quality of organic amendments in 2012 in Saria, Burkina Faso.

Treatment	Plant P uptake	Partial factor productivity	P partial balance
	kg ha^−1	kg kg^−1	kg ha^−1
Control	1.15	18.0 (1.08)	11.95
Straw	2.93	36.6 (1.55)	12.47
Manure	9.96	50.7 (1.67)	26.54
Compost	5.81	41.7 (1.58)	22.69
p values	<0.001	0.032	<0.001
Lsd (0.05)	2.22	0.45	2.22
CV (%)	36.4	24.7	9.8

Values in the brackets are log transformed.

Sorghum P uptake varied from 1.15 (control treatment) to 9.96 kg ha⁻¹ (manure treatment). The manure treated plots recorded the highest sorghum P uptake followed by the compost treated plots. The differences in sorghum P uptake in the control and sorghum straw treatments were not significant at p < 0.05.

The partial factor productivity of P varied from 18 kg grain per kg of P in the control treatment to 50.7 kg grain per kg of P in the manure treatment. All the three types of organic amendments increased significantly the partial factor productivity compared to the control but there were not significant differences among the organic amendments treatments in increasing the partial factor productivity of P.

The P partial balance was positive in all treatments and varied from 11.95 kg ha⁻¹ in the control treatment to 26.54 kg ha⁻¹ in the manure treatment and ranked as follows: control = Straw < compost < manure.

Relationships between sorghum yield and P use efficiency and soil chemical properties

Figure 2 shows the results of the Principal Component Analysis (PCA) between sorghum grain yield P partial productivity and soil properties. The two Principal Components axes (F1 and F2) explained 88.5% of the total variability. The first Principal Component axis (F1) accounted for 71.8% of the total variability among the different types of organic amendments and the second (F2), 16.7% of the total variation. Sorghum grain yield and the partial factor productivity of P were significantly correlated to soil pH, carbon, total nitrogen and organic P. The different types of organic amendments were mainly distributed along the first axis in the order manure ≥ compost > sorghum straw > control.

Figure 2. Principal Component Analysis of sorghum yield, P use efficiency and different soil properties of long term experiment in Saria, Burkina Faso in 2012. G-Y: grain yield, Res-P: resin-Pi, PFP: partial factor productivity of P.

Discussion

Change in soil properties

The higher values of soil pH, organic carbon and total nitrogen observed in treatments receiving manure, compost and straw over the control treatment are in agreement with earlier reports on the ability of organic amendments to improve soil properties (Agbenin and Goladi 1997; Mokolobate and Haynes 2002, Koulibaly et al. 2017). The differences in the effect of different types of organic amendment on soil pH and organic carbon were obvious only 32 years after the start of the experiment and demonstrates the importance of long term studies on soil properties that are needed to make informed decisions on soil fertility management options. The fact that manure and compost significantly increased soil carbon and soil pH than sorghum straw 32 years after the start of the experiment can be partly attributed to their chemical characteristics. The pH of manure and compost are indeed higher than that of sorghum straw (Table 1). The low C/N ratio of manure and compost was probably more favourable for carbon accumulation in soil than sorghum straw. Samahadthai et al. (2010) reported that the application of organic amendment with low C/N ratio promotes the accumulation of organic carbon in soil. Sorghum straw also content more easily decomposable organic material than manure and compost (Hien et al. 2004) which can explain its low effect in increasing soil carbon with reference to manure and compost. The decrease in soil pH, organic carbon and nitrogen content in all treatments receiving organic amendments from 10 years after the start of the experiment to 32 years after the start of the experiment suggest that the rate of 10 t ha⁻¹ of organic amendment combined with mineral fertiliser is not sufficient to maintain soil quality. Koulibaly et al. (2017) have also noted a decrease in soil organic carbon, nitrogen and pH in a long term cultivated Lixisol where maize straw is annually incorporated with compost or farmyard manure. The decrease of soil organic carbon and nitrogen in such type of soils can be partly attributed to the high turnover of organic materials in these soils exacerbated by the annual tilling (Bationo and Buerkert 2001). These results reinforce the need for an integrated soil fertility management options implying the use of both mineral and organic fertilisers, the use of nitrogen fixing crops in a rotation system and the adoption of minimum soil disturbance technologies.

Soil P forms and accumulation

The three organic amendments used in this study differed in their total P content (Table 1) which could explain their varying effects in increasing soil total P. It was indeed observed that the higher the organic amendment P content the higher its effect in increasing soil total P. The increase in soil total P and inorganic P pools from 10 years after the start of the experiment to 32 years after the start of the experiment could be partially attributed to the low P uptake by sorghum (less than 30% of the total amount of P applied) which resulted to a positive P balance (Table 5). Increases of soil organic phosphorus have often been linked to the increase of soil organic carbon (Bünemann et al., 2006; Hou et al., 2014) which may explain the higher organic phosphorus content in the manure and compost treatments over that of the straw and control treatments. However, it seems that this effect of organic amendments in increasing soil organic phosphorus occurred only during the first years of cultivation of sorghum since soil organic phosphorus remained stable from year 1990 (10 years after cultivation) to year 2012 (32 years after cultivation). Other studies have shown no or little effect of organic amendments on soil organic phosphorus and no relationship has been established between soil organic phosphorus and soil organic carbon in long-term field trials (Keller et al. 2012; Annaheim et al. 2015). These results show that beyond soil organic carbon and the age of soil use other factors govern the turnover of soil organic phosphorus. The higher values of soil Resin-P measured in the manure and compost treatments with reference to the control treatment corroborate with earlier reports on the effect of organic amendments in increasing available P pools (Lompo et al. 2008; Shafqat and Pierzynski 2010). The positive effect of manure and compost on soil Resin-P could be attributed to their high P content in available forms. The low carbon to phosphorus (C/P) ratio of these organic amendments which is less than 200 (Table 1) are indeed favourable for P release in soil (Kwabiah et al. 2003). Furthermore, many studies have shown that animal manure contains a certain amount of P in Resin extractable P forms (Turner and Leytem 2004; Takahashi 2013). The positive effect of manure and compost on soil Resin-P can also be attributed to the capacity of organic matter to reduce P fixation by the formation of bonds with some cations such as Fe and Al (Marschner and Rengel 2012). The low capacity of sorghum straw to increase soil Resin-P may be partially attributed to its low P content and high C/P ratio of 1884. Other studies have also found little effect of organic amendments with low P content on soil readily available P pools (Verma and Penfold 2017).The fact that manure initially contain some amounts of NaHCO₃-Pi and NaOH-Pi forms, may explain its effect in increasing soil NaOH-Pi and NaHCO₃-Pi (Turner and Leytem 2004; Takahashi 2013). Furthermore, several studies have showed that residual fertiliser P is found predominantly in labile and moderately labile form (Zheng et al. 2002; Zhang et al. 2004) which could explain the high effect of manure which have the higher P content, in increasing soil NaHCO₃-Pi and NaOH-Pi over compost and sorghum straw. NaHCO₃-Pi and NaOH-Pi are respectively labile and moderately labile phosphorus which can contribute to plant nutrition (Zheng and Zhang 2012). In this study, no evidence of the contribution of NaHCO₃-Pi and NaOH-Pi to plant nutrition has been shown as they increase with year.

HCl-P has been found to be a stable P fraction which is not affected by soil fertility management during long term experiments (Zhang et al. 2004; Keller et al. 2012). However, organic amendments usually contain P in HCl-P which may justify the increase of soil HCl-P in compost and manure treatments in this studies 32 years after the start of the experiment (Turner and Leytem 2004; Takahashi 2013). Sharpley et al. (2004) have also found in a long-term experiment that this form of P is increased after addition of different types of manure. HCl-P is an apatite P which can also be solubilised when soil pH become very acidic (Zapata and Roy 2004), which may explain the decrease of HCl-P in the control treatment.

Change in Sorghum yields, P use efficiency and their relations to soil characteristics

The fluctuation in sorghum grain yield as shown in Figure 1 may be explained partly by both the variation of rainfall amount and the change of sorghum varieties which differ either by their photosensitivity or by their yield potential from the beginning of the experiment to 32 years after the start of the experiment. By grouping sorghum grain yield by time frame of about ten years we noted a change on the effect of sorghum straw in increasing sorghum grain yield which was significant only in the first decade compared to the control treatment. Little effect of sorghum straw on sorghum yield after 26 years of application has also been reported by Diallo-Diagne et al. (2016) working in the same experiment. As mineral fertilisers are yearly applied in all treatment, the decrease in the effect of sorghum straw could be partly linked to the high decrease in soil organic carbon and soil pH noticed in this treatment. This hypothesis is further sustained by the PCA results which show a significant correlation between soil pH and organic carbon and yield while the correlation between grain yield and soil resin P was not significant. In the same view we can explain the positive effect of manure and compost in increasing sorghum yield during all the studied time frames by the fact that these organic amendments tend to maintain soil carbon and decrease soil acidification. Soil acidity can determine crop production by reducing nutrients availability and uptake through the release of ions such as aluminium which can hinder roots nutrients uptake or ions such as iron which can complex phosphorus (George et al. 2012). Likewise, organic carbon can favour crop growth by the increase of nutrient availability (Marschner and Rengel 2012). The P uptake was indeed lower in the control treatment and sorghum straw treatments where soil pH and organic carbon are low and higher in compost and manure treatments where soil pH and organic carbon are high (Table 5). The higher partial factor productivity of P obtained from the application of organic amendments relative to the control treatment can be explained by the fact that organic amendments tend to improve soil properties (the increase of soil pH, soil carbon and nitrogen content), which favour crop P uptake and growth. However, although manure and compost showed higher effectiveness in increasing soil fertility and accumulating P in soil than sorghum straw, they induced similar partial productivity of P as sorghum straw. Furthermore the values of partial productivity of P obtained with these organic amendments remains low when compared to the optimum values of partial productivity of P which are between 100 and 250 kg grain per kg of P applied (Fixen et al. 2015). These low rates of partial P productivity obtained with compost and manure might be due to the fact that these organic amendments when combined with mineral fertilisers provide high quantity of P that may not be necessary used for biomass production. It has indeed been showed that P use efficiency becomes low when the amount of P applied gets high (Nziguheba et al. 2016). Hence, the amount of mineral P fertilisers applied must be reduced when such organic amendments with high available P content are combined.

In conclusion, the effect of long-term use of organic amendments on soils P forms is dependent on their quality. We observed a significant increase in soil available and unavailable P with the application of organic amendment with high P content and low C/P ratio. The high accumulation of P throughout years in labile forms in soil treated with certain quality of organic amendments such as manure and compost which does not reflect the productivity of applied P should be a justification for recommending new formulation of mineral fertiliser application rates when those organic amendments are added. The choice of organic amendments for optimising sorghum P uptake and yield in long term must be driven by their capacity and potential for maintaining soil carbon and limiting soil acidification.

Acknowledgments

We thank the Alliance for Green Revolution in Africa (AGRA) for the PhD grant to the first author and the International Foundation for Science (IFS) for partly financing this work through grant No C/4918-1.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributors

Dr Dohan Mariam Soma, PhD, is a soil scientist at the Institut de l'Environnement et de Recherches Agricoles in Burkina Faso. Her research works focus on the sustainability of soil fertility management options through long‐term studies and the use of agro minerals such as phosphate rock to improve soil phosphorus availability. She is presently involved in a collaborative research project on the establishment of the model for fertilizing cultivation promotion using Burkina Faso phosphate rock.

Dr Delwendé Innocent Kiba, PhD, is a soil scientist at the Institut de l'Environnement et de Recherches Agricoles in Burkina Faso. He is interested in research for development related to integrated soil fertility management and food security in Africa. He is currently coordinating a project on the development of sustainable soil management options in yam systems in West Africa.

Dr Nana Ewusi‐Mensah, PhD, is a lecturer at the Department of Crop and Soil Sciences. Kwamé Nkrumah University of Science and Technology in Ghana. His research works focus on soil biology, composting organic amendments, biochar, and integrated soil fertility management. He is presently involved in a research project on Arbuscular mycorrhizal fungi inoculation on soybean (Glycine max) productivity in the semi‐deciduous forest zone of Ghana.

Dr Zacharia Gnankambary holds a PhD in soil sciences from Swedish University of Agricultural Sciences (SLU), Umeå, Sweden. He is working as a soil scientist at the Institut de l'Environnement et de Recherches Agricoles in Burkina Faso and focuses on agroforestry, soil chemistry and soil microbiology. He is the former director of the Bureau National des sols (BUNASOLS) and former chairman of the Board of the phosphate mining society in Burkina.

Dr François Lompo is a soil scientist at the Institut de l'Environnement et de Recherches Agricoles in Burkina Faso. He has conducted numerous research studies on the use of phosphate rock and organic amendments in agriculture. He has been the director of INERA from 2011 to 2014 and the Minister of Agriculture, Water Resources, Sanitation and Food Security of Burkina Faso from 2014 to 2015. He was in 2014 elected Chairman of the Board of Directors of the West and Central African Council for Agricultural Research and Development (CORAF/WECARD).

Dr Michel Papoaba Sedogo is a former soil scientist at the Institut de l'Environnement et de Recherches Agricoles in Burkina Faso. He is a pioneer researcher in soil science in Burkina Faso. He has led numerous research projects and supervised numerous PhDs. He is retired and actually member of the National Academy of Sciences, Arts and Letters of Burkina Faso (ANSAL‐BF).

Prof Robert Clement Abaidoo, PhD, is a professor of soil science at the Kwame Nkrumah University of Science and Technology (KNUST) in Ghana and soil microbiologist at the International Institute of Tropical Agriculture (IITA) in Nigeria. He has led, coordinated or managed over 26 research/development projects and is involved in the Steering Committee Chairmanship of DANIDA/ BSU (Danish International Development Agency/Building Stronger Universities) Project. He has led the AGRA‐KNUST Soil Science PhD study program for West Africa at KNUST (2010) which partly funded this study.