Introduction

Arid and semiarid agroecosystems represent about 41% of the total terrestrial surface on global scale, and more than 35% of the world population depend on these dryland areas for their food (Pointing and Belnap 2012). Primary production of these dryland ecosystems is generally limited by the availability of water and nitrogen (N) (Ehleringer 2001; Schwinning et al. 2005). However, accumulation of N as nitrate (NO3−) has been reported in many laboratory and field investigations of dryland ecosystems, despite the low intrinsic soil N contents (Cookson et al. 2006). Grasslands are important terrestrial ecosystems which store carbon (C) and N on a long-term basis and provide the additional service of soil and water conservation, especially in semiarid regions (Silver et al. 2010; Gao and Cheng 2013).

Atmospheric N deposition has increased manyfold from 15 Tg N/year in 1860 to 187.15 Tg N/year in 2005, and this rate of N deposition is expected to double in the next 25 years (Galloway et al. 2004; Li et al. 2015). Literature reviews highlighted the impact of N pollution from atmospheric deposition on grassland biodiversity (Vitousek et al. 1997; Stevens et al. 2004; Sinsabaugh et al. 2015). Increasing levels of N deposition and pollution from the atmosphere onto seminatural grasslands disturb the natural N cycle and may ultimately result in nitrogen saturation (Reynolds and Edwards 1995). The deposition of N species from the atmosphere onto seminatural ecosystems has the potential to induce N leaching and change plant biodiversity and, hence, pose a global concern (Phoenix et al. 2006, 2012).

Previously, it was believed that soil pollution and nitrate pollution of surface waters and groundwaters predominantly were problems of developed countries (Follett 1992); however, the developing countries are facing similar problems now. In developing countries, injudicious use of N fertilizers for increasing crop production, coupled with irrigation, is quite common and may be a major cause of groundwater pollution. Nitrate leaching is potentially quite a serious problem in these areas (Bijay et al. 1995). Compared to developed countries, in terms of evaluating the potential pollution risk of N pollution, the developing countries have two advantages: one is the lower availability of resources like capital and expensive fertilizers; the other is that the trend to use ever-increasing amounts of synthetic fertilizer to increase crop yields came quite late there compared to its start in developed countries.

According to Tahir and Rasheed (2008), 13% of total soil contamination problems are nitrate-based, 68% bacterial, 24% arsenic-based, and 5% from fluoride pollution. They further suggest that nitrate N is mainly responsible for soil N pollution problems and ammonium N is not a major concern. It has been commonly believed that arid and semiarid ecosystems have low N availability, low inputs via atmospheric deposition and N2 fixation, and high rates of gaseous losses of NOx, NxO, N2, and NH3 (Peterjohn and Schlesinger 1990). This is certainly strongly believed for the area around Faisalabad because of the dry and hot weather and low rainfall. However, many studies over recent decades have demonstrated that net N mineralization increases with wetting of dry soil (Cui and Caldwell 1997; Saetre and Stark 2005). Therefore, irrigation during hot periods and little precipitation can trigger mineralization pulses and is the cause of high N pollution of drainage waters even during dry and warm periods. Work on other ecosystems also showed that rainfall affects mineralization, N availability, soil moisture, and plant water use efficiency (Hooper and Johnson 1999; Austin et al. 2004; Chen et al. 2005).

Soil properties and climatic conditions greatly affect nitrate leaching and accumulation in soil (Macduff et al. 1990a, b). In Pakistan, most research is based on cultivated areas and nitrate-N leaching in response to pollution. The pH of Faisalabad’s soils is generally in the range 7.3–8.2 (Hussain et al. 2010), which is quite favorable for nitrification and nitrate-N leaching. Faisalabad’s soils include course-textured sandy loams, and some such soils may have less ammonium-N retention capacity compared to fine-texture silty clays (Akhtar and Memon 2009). Fortnightly irrigation during dry hot weather is essential for vegetation survival in some places. So, in this area, both nitrate-N and ammonium-N leaching may result from soil pollution.

There is clearly a risk that, over time, levels of N in soil are increasing with increasing fertilizer inputs. Nitrogen fertilizer use in Pakistan was one million nutrient tonnes in 1980–1981, two million tonnes in 1992–1993, and three million tonnes in 2002–2003 (FAO 2007). Because the area surrounding Faisalabad is agricultural and canal water and tube well water at pH 7.48–8.0 are used for irrigation (Hussain et al. 2010), approximately 31% of the area is contaminated by high levels of nitrate-N (Tahir and Rasheed 2008) and ammonium-N. Irrigation of grasslands, even without adding any further fertilizer, could be the cause of soil and groundwater pollution, to an extent likely to depend upon soil organic matter content, C:N ratio, pH and climatic factors such as rainfall pattern (intensity and temporal distribution), and temperature. Therefore, quantitative assessment of N cycling in soil/grassland systems in Pakistan was deemed by the authors to be very important for understanding N species transformation mechanisms and assessing the amounts of nitrate-N and ammonium-N needed for sustainable production in grassland areas of Pakistan and any potential associated N pollution problems. The present study aims to provide an assessment of N cycling within selected soils throughout the year and to assess seasonal variations in N species transformations with particular reference to assessing risk of leaching losses.

A recent study of N cycling under grassland in York, UK, demonstrated a close match between the dynamics of mineral-N production in vegetation-free soil and those of plant N uptake requirements (Bhatti et al. 2013). It was suggested that this is a consequence of natural evolution for N conservation under less polluted conditions. We hypothesize that quantifying the rates of production of ammonium-N and nitrate-N in vegetation-free soils near Faisalabad throughout the year will show that potential for ammonium-N and nitrate-N leaching during winter, attributable to mineralization in soil, will be reduced by lower mineralization of senescent material deposited as litter in winter, as in the UK, and immobilization of mineral-N inputs outside the period of grass growth. If so, atmospheric N inputs, rather than the soil-generated mineral-N, will largely regulate water-soluble mineral-N concentrations in the soils. A second hypothesis is that, because of the lower extent of past use of fertilizers and of atmospheric N pollution in Pakistan, soils in Pakistan are likely to be more conservative of N inputs compared to the soils in the UK.

Materials and methods

Study domain, description of study site, and meteorological conditions

Pakistan is a developing country with an economy based on agriculture. Its total area is 80.5 Mha of which 20.36 Mha is cultivated. About 76.7% (16.23 Mha) of that area is irrigated by canals or tube wells, and only 20.3% (4.13 Mha) is dependent on rainfall (Ayub et al. 1999). The climate of Faisalabad in Punjab province is hot and dry, but semihumid during the rainy season. Average annual rainfall in the region is 372 mm, and majority of this rainfall occurs during the monsoon period of July–September (Farhat et al. 2014).

Soil samples were collected from January 2010 to December 2010 (at approximately 1-month intervals) from a transect at the Sports Field at the University of Agriculture, Faisalabad, Pakistan (31.4310° N, 73.0695° E; elevation 183 m) in an area remote from human and animal activities. It is also a regularly mown area dominated by Bermuda grass. Its topography was almost level and the soil was visibly uniform.

Soil sampling and preparation for analysis

Samples were collected monthly from mid-January until December. Using a soil augur, at each sampling site, six bulk field replicates were collected along a transect at the Sports Field, University of Agriculture, Faisalabad, Pakistan. Each of the six field replicates contained soils from a minimum of five cores, combined to give the six bulk samples. Samples were taken from 0–10 and 10–20 cm depths separately. Stones, grass leaves, and roots were removed as much as possible during collection, and collected samples were stored at 2–4 °C in sealed, clearly labeled, polythene bags within an hour of collection.

In the laboratory, after any remaining visible plant fragments and small pebbles were removed, each soil was mixed thoroughly, and subsamples were taken from each field replicate sample for analysis. The remaining soil was used for the incubation experiment.

Determination of soil pH, moisture content, and texture

Soil pH was measured using a glass/calomel electrode and pre-calibrated digital pH meter (Fisher brand, Hydrus 400) in a soil:water (1:5) mixture, after shaking for 30 min and allowing the mixture to equilibrate at room temperature (Rayment and Higginson 1992).

On the same day of sampling, for measuring soil moisture contents, 10 ± 0.01 g subsamples of soil from each replicate were weighed into pre-weighed aluminum foil dishes, dried for 24 h at 105 °C in a pre-heated oven, and then cooled in desiccators. After drying, the dishes of dry soil were reweighed and the difference in weight was recorded to determine the moisture contents from the loss of mass, expressed on an oven dry weight basis (Riaz et al. 2008).

Soil textural analysis was performed only once on the soil samples from January sampling at the start of the experiment using the Bouyoucos hydrometer method of Gee and Bauder (1986).

Soil incubation under laboratory conditions

Subsamples of field-moist soil (10.0 ± 0.01 g) were incubated in a well-protected structure adjacent to the author’s experimental site in wide-mouth, loosely capped glass jars (to facilitate air entry but to exclude rainfall and plant uptake) for 7 days (Riaz et al. 2008). A pre-incubation period of 7 days was chosen to allow adequate time for the conditions and N mineralization/microbial immobilization to re-establish for plant-free soil. This was necessary to have uniformly moist soil with homogenous conditions before the start of the experiment (Riaz et al. 2009). It is accepted that this is somewhat arbitrary, and microbial activity in bulk soil could differ from that occurring in the rhizosphere under field conditions. However, this was deemed acceptable because the authors were primarily interested in variations in N speciation between months and any such variation should clearly show up. It was important to quantify the capacity of vegetation-free soil to produce mineral-N species.

The amounts of KCl-extractable ammonium-N and nitrate-N in pre- and post-incubated soils were extracted with 1.0 M KCl at 1:5 (w/v basis). The mixture was shaken for 90 min at 150 rpm on an orbital shaker and filtered through Whatmann#42 filter papers. The resultant samples were immediately stored at 4 °C before analysis for ammonium-N and nitrate-N.

Net mineral-N production for each soil layer was calculated using the following expression:

$$ \mathrm{Net}\ \mathrm{mineral}-\mathrm{N}\ \left(\mathrm{mg}/\mathrm{kg}\ \mathrm{soil}\right)=\left(\mathrm{post}-\mathrm{incubated}\ \mathrm{ammonium}-\mathrm{N}+\mathrm{post}-\mathrm{incubated}\ \mathrm{nitrate}-\mathrm{N}\right)-\left(\mathrm{pre}-\mathrm{incubated}\ \mathrm{ammonium}-\mathrm{N}+\mathrm{pre}-\mathrm{incubated}\ \mathrm{nitrate}-\mathrm{N}\right) $$

Soil chemical analyses

Ammonium-N and nitrate-N plus nitrite in the extracts were determined by steam distillation of ammonia (NH3), using heavy MgO for ammonium-N and Devarda’s alloy for nitrate-N analysis. The distillate was collected in saturated H3BO3 and titrated against dilute H2SO4 at pH 5.0. The KCl-extractable mineral-N was calculated by adding ammonium-N and nitrate-N using the method described by Buresh et al. (1982) and Keeney and Nelson (1982). Soil organic C contents were measured by dichromate oxidation and subsequent titration with acidified ferrous ammonium sulfate using the Walkley and Black method (Walkley and Black 1934). Soil total N was determined by the method of Bremner and Tabatabai (1972) following Kjeldahl’s digestion and distillation procedure.

Rainwater analysis

Rainfall is not frequent under Faisalabad conditions in Pakistan; therefore, samples were collected for chemical analysis only in wet months by using duplicate rain gauges installed at the sampling site. Rain samples were analyzed for nitrate-N and ammonium-N contents using the method described above.

Statistical analysis

One-way analysis of variance (ANOVA) test was used to study the difference in soil variables between the months, and Tukey’s post hoc test was used to compare the multiple means when difference was significance between the sampling months. Independent-samples test was used to find whether there was significant difference between the pre-incubated and post-incubated mineral-N contents. Tables and figures always contain means and standard error of means from six field replicates unless otherwise stated. All statistical analyses were performed using SPSS version 19.

Results

Soil properties and environmental conditions

Soil used in the study was clay loam in textures at both depths. The pH of the water of soil from 0–10 and 10–20 cm depths was 7.38 and 7.50 respectively. Table 1 shows how temperature and relative humidity at the site varied throughout the year. May and June were the warmest months. Relative humidity (RH) ranged from 30.90% in May to 82.32% in January.

Table 1 Changes in air temperature and relative humidity (RH) at the site throughout the study period. The values shown are the means over each month
Full size table

Ammonium-N and nitrate-N concentrations at 0–10 cm depth

The seasonal trends in ammonium-N and nitrate-N concentrations in fresh, field-moist soil from 0–10 cm depth pre-incubation and in soil post 7-day incubation at ambient outdoor temperature are shown in Fig. 1a, c. The pre-incubation data in Fig. 1a, therefore, show how the extractable soil ammonium-N concentration in soil from the top 10-cm layer of the transects at Sports Field, expressed in units of milligrams N per kilogram, changed seasonally.

Fig. 1
figure 1

Changes over the sampling year in the concentrations of ammonium-N (mg/kg soil) in soils from a 0–10 cm and b 10–20 cm and nitrate-N (mg/kg soil) in soil from c 0–10 cm and d 10–20 cm depths from the Sports Field pre-incubation and after incubation for 7 days at ambient outdoor temperatures in Faisalabad (post-incubation). Results are means of six replicates and contain the ± standard error of the means. Where pairs of bars have an asterisk, the effect of incubation is significant at P < 0.05

Full size image

There were also significant differences (P < 0.05) for the ammonium-N concentrations at 0–10 cm depth between months over the study period (Table 2). Pre-incubation data show that the ammonium-N concentration in January differed significantly from that in other months except February, April, and September (Fig. 1a; Table 2). The March value was significantly higher than the values in all other months. The concentration in April was the highest and differed significantly from that in other months except January, August, September, and October. The minimum June value differed significantly from values for all months except May. Overall, there is clear evidence of low summer values, an increase in autumn and early winter, and a peak in March. Post-incubation data for ammonium-N at 0–10 cm depth showed the concentrations in January and April did not differ significantly from each other, but they differed significantly from values in other months (Fig. 1a; Table 2). Concentration in June was significantly lower than that in all other months. The highest concentrations of ammonium-N in pre- and post-incubated soils were observed in March whereas values were least in the June soil samples of the upper layer. The concentration increased in spring with increasing temperature, but in summer, the concentration was much lower at high temperature.

Table 2 Significant differences between months in concentrations of ammonium-N, nitrate-N, and net mineral-N in soil, pre-incubation and post-incubation from 0–10 and 10–20 cm depths
Full size table

Fresh samples tended to show slightly higher ammonium-N concentrations than the post-incubated samples except in January, April, May, and August (Fig. 1a). In the post-incubated samples, the relatively small decrease in ammonium-N concentration indicated the net effect of immobilization, volatilization, and conversion of ammonium-N to nitrate-N. All months except May, June, and August showed significant difference in ammonium-N concentration before and after incubation. A general trend of relatively higher levels of ammonium-N in winter than in summer was observed.

Figure 1c shows nitrate-N concentration over the year in turf soil from 0–10 cm from January to December. Pre-incubation data showed that nitrate-N concentration in January differed significantly from that for all other months except March and June (Fig. 1c; Table 2). The concentration in February was significantly higher than that in all other months. The March values differed significantly from values for the other months except January, June, and October. The pre-incubated nitrate-N concentrations for the months of May, July, August, September, October, November, and December remain more or less similar. Post-incubation data show that nitrate-N in soil from 0–10 cm in January differed significantly from values for the other months except May, June, October, and December (Fig. 1c; Table 2). The February values differed significantly from values for other months except July and September. March and April values significantly differed from each other and from values for all other months.

In pre-incubated fresh soil samples, the highest concentrations of nitrate-N were observed in February whereas the minimum concentrations of nitrate-N were present in April. In post-incubated soil samples, the highest concentration was observed in March and the minimum in April. The concentrations of nitrate-N decreased with increasing temperature in spring, and when temperature was high in summer, the concentration was lower compared to that in cooler months.

Throughout the year except March, fresh samples generally showed slightly higher nitrate-N concentrations compared to post-incubated samples (Fig. 1c). During March, pre-incubated soils showed low concentrations of nitrate-N compared to those of the incubated samples. Therefore, nitrification possibly accounted for most of the decrease in ammonium-N in March. All months except January, April, June, July, and September showed significant differences in nitrate-N concentrations after incubation. The low April values suggest nitrate uptake by grass during the little bit of March rainfall.

Ammonium- and nitrate-N concentrations at 10–20 cm depth

The seasonal trends in ammonium-N and nitrate-N concentrations in fresh, field-moist soil from 10–20 cm depth pre-incubation and in soil post 7-day incubation at ambient outdoor temperature are shown in Fig. 1b, d. The pre-incubation data in Fig. 1b therefore show, for example, how the extractable soil ammonium-N concentration in soil from the 10–20 cm layer, in units of milligrams N per kilogram, changed seasonally. Figure 1b, d also shows how the ammonium-N concentrations and nitrate-N concentrations in soil from 10–20 cm changed between each sampling and, subsequently, as a consequence of incubation for 7 days at ambient outdoor temperatures. Table 2 shows the significant differences (P < 0.05) between months all through the year.

Pre-incubation data show that ammonium-N concentrations in January differed significantly from those in other months except February and April (Fig. 1b; Table 2). August, September, and October values were not significantly different from each other, but they differed significantly from values for all other months except April. The November and December values were not significantly different from each other, but they differed significantly from values for other months. Post-incubation data showed ammonium-N concentrations in January differed significantly from values for other months except April, September, and December. The April value differed significantly from values for all months except January.

For pre- and post-incubation data, the highest ammonium concentrations were observed in March whereas the lowest were in June (Fig. 1b). When considered in the context of temperature, the ammonium-N concentration increased with the temperature boost in spring but in early summer it decreased to a significant extent. Temporal trends are similar to those for the 0–10-cm soil, probably reflecting mobility of ammonium-N and nitrate-N between the 0–10- and 10–20-cm soil layers. Through much of the year, samples showed relatively higher ammonium-N concentration compared to post-incubated samples, except for January, April, May, and August. However, in April and May, differences after incubation were not statistically significant (Fig. 1b).

Figure 1d shows how nitrate-N concentrations varied in turf soil from 10–20 cm from January to December. Pre-incubation data showed nitrate-N concentrations in January differed significantly from those in other months except May, August, September, October, and December (Fig. 1d; Table 2). The highest values in February differed significantly from values for all other months except July. Concentrations in July differed significantly from those for all months except February. Post-incubation data showed that nitrate-N concentrations in January, May, October, and December did not differ significantly from each other, but they differed significantly from those for other months (Fig. 1d; Table 2).

In pre-incubated soil samples, the highest nitrate-N concentrations in the 10–20-cm soils were in February whereas the lowest concentration was in April. For post-incubated soil samples, the peak values were in July, while the minimum concentration was in April (Fig. 1d). The fresh samples show relatively higher nitrate-N concentrations compared to post-incubated samples over October, November, December, January, February, and also in May and August but lower in March, April, June, July, and September. All months except July showed significant differences between soils before and after incubation. Both 0–10- and 10–20-cm soils show similar trends. These appear to show a clear trend towards nitrate-N immobilization during incubation from October to February, bearing in mind the apparent ammonium-N immobilization or loss by volatilization and nitrification from September to December.

Soils from 0–10 cm and from 10–20 cm show quite similar seasonal trends. Good correlations were observed between ammonium-N values at 0–10 and 10–20 cm and between nitrate-N values at 0–10 and 10–20 cm before and after incubation (data not shown). Ammonium-N values showed a strong correlation pre-incubation (R2 = 0.986) with a slope of 0.827; after incubation, the correlation was still strong (R2 = 0.974) but the slope decreased to 0.771, suggesting loss of any mobility effect between soil layers. The strong correlation between concentrations at both depths suggest mobilization between 0–10 and 10–20 cm depths under field conditions. Nitrate-N values also showed a strong correlation between depths (R2 = 0.897, slope 1.071) before incubation compared to after incubation (R2 = 0.735, slope 1.09).

Changes in net mineral-N concentrations at 0–10 and 10–20 cm depths

Figure 2 shows how concentration of net mineral-N changed as a result of incubation in soil from 0–10 and 10–20 cm over the 12 months. Table 2 shows the significant differences (P < 0.05) between net mineral-N over the 12 months. Data for soil from 0–10 cm depth show that change in mineral-N in January was significantly smaller than that in February, March, October, November, and December. That in May was significantly smaller than those in February, October, November, and December (Fig. 2; Table 2). The November value was significantly higher than values for all months except December (i.e., gave the greatest net decrease). The low changes in the concentration of mineral-N in January during incubation of soil from 10–20 cm depth (Fig. 2) differed significantly from values for all months except April, May, July, and August. The small increase after incubation in April differed significantly from values for February, March, October, November, and December. That for August differed significantly from those for February, March, October, November, and December.

Fig. 2
figure 2

Changes over the 12 months in the concentrations of net mineral-N in soil from 0–10 and 10–20 cm during incubation for 7 days at ambient outdoor temperatures in Faisalabad. Results are means of six replicates and contain the ± standard error of the means. Where pairs of bars have an asterisk, the effect of sampling depth is significant at P < 0.05

Full size image

It is evident from Fig. 2 that, for both soil depths, there was net reduction in mineral-N production (immobilization) after incubation and this trend was dominant over the entire sampling period with few exceptions. Net mineral-N was significantly different between the sampling depths for the months of January, June, and July.

Changes in total soil C and N contents

Table 3 shows the seasonal trends for soil C and N concentrations (expressed in percentage) throughout the year from 0–10 and 10–20 cm depths. Carbon percentage in the soil at both depths shows a similar seasonal trend. Soil C contents were the highest in May in soil from 0–10 and 10–20 cm depths. Non-significant differences were observed during the whole sampling year at the P < 0.05 level between soils from 0–10 and 10–20 cm depths for soil C contents.

Table 3 Changes over the sampling year in the percentage of C and N in soil from 0–10 and 10–20 cm from the University Sports Field
Full size table

Nitrogen percentage shows a relatively clearer trend compared to soil organic C contents, with higher N concentrations during winter months (November to January) and generally lower N percentage over spring to summer. Mostly the 0–10-cm soil showed slightly higher N percentage compared to the 10–20-cm soil. However, significant differences at the P < 0.05 level were observed between 0–10 and 10–20 cm only during the months of November and December.

Rainfall N concentrations and moisture contents in grassland soils

The rainfall intensity and N contents of rainfall are shown in Fig. 3. Rainfall was light in the study area for most of the year. Substantial rain only occurred in July, August, and September with more than half of the rainfall occurring in the monsoon, and rainfall in other months was low and not predictable (Fig. 3a). Light rainfall was also recorded in other months, however. The analysis did not reveal any considerable input of nitrates and ammonium via rain in the driest months. Clearly though, rainfall from July to September contributed towards nitrate and ammonium leaching to deeper soil horizons below those analyzed here, bearing in mind the concentrations of extracted ammonium-N and nitrate-N (Fig. 3b). Figure 4 shows how soil moisture content varied at time of sampling over the year. Under the influence of rain, significantly (P < 0.05) higher soil moisture contents were observed at 0–10 and 10–20 cm depths during the August sampling. Soil moisture contents were also higher for both soil layers during the months of January and February.

Fig. 3
figure 3

Monthly changes in a the amount of rainfall (ml/m2) and b the amount of nitrate-N and ammonium-N (mg/m2) from January 2010 to December 2010

Full size image
Fig. 4
figure 4

Changes in soil moisture contents (%) over the 12-month period from January 2010 to December 2010 at 0–10 and 10–20 cm depths. Values are means of six replicates and contain ± standard errors of the means. Black bars with different small letters indicate significant soil moisture content differences at 0–10 cm depth, and grey bars with different capital letters show that soil moisture content was significantly different at 10–20 cm depth over the 12 sampling months

Full size image

Discussion

Comparison to the UK sites

From the studies described by Bhatti et al. (2013), it was deduced that there are strong seasonal trends in N cycling in soils under grassland in the UK, driven primarily by variations in temperature and growth of grass over the year. In mid-summer, growth is vigorous and return of organic matter from litter fall to the soils occurs at a slower rate than at other times of the year. Therefore, there is little fresh litter subject to mineralization at this time of year, a time when readily decomposable litter from the previous winter and early spring has already decomposed to a large extent. Thus, it was concluded that mineral-N concentrations in fresh soils and associated drainage waters fall only partly because of the long-recognized plant uptake effect in summer; much of the decline is attributable to the low mineralization rate resulting from low supply of readily mineralizable organic matter in summer months. This was discernible only because Bhatti et al. (2013) opted to study seasonal changes in available mineral-N pools in fresh, field-moist soils and mineralizable soil N in soils at ambient outdoor temperatures after removal from living plant biomass. That approach was therefore adopted in the present study.

In Pakistan, the climatic conditions are very different from those in the UK. We recorded a very uneven distribution of rainfall throughout the year at the test site, with heavy rainfall only in July, August, and, to a lesser extent, September. January to June and October to December are very dry. When temperature started to rise in April, the relative humidity falls to < 40% and soil moisture content at the time of sampling from May to July and from September to November was very low.

Seasonal variations in N species transformations

The data show quite substantial seasonal changes in the soil C and N contents at both 0–10 and 10–20 cm depths. Soil C contents fell markedly in March and April and remained lower in July through October. Interestingly, the soil N contents started decreasing from February to April and showed a clear trend towards higher N contents during winter months compared to the rest of the year at both depths.

Although there was a continuous increase in mean daily temperature from January (11.1 °C) to February (15.7 °C) and March (23.5 °C) and little rainfall during March and low relative humidity, maximum oxidation of organic matter was observed. Despite plant uptake and volatilization losses, high concentrations of extractable ammonium-N and nitrate-N at both 0–10 and 10–20 cm depths during March show the oxidation rate is quite high. No mineral-N was added via irrigation water in March.

During March, the level of nitrate-N at both 0–10 and 10–20 cm also increased in post-incubated samples, showing the loss of ammonium-N during incubation was largely because of nitrification. It is probable that nitrification starts to build up slowly on rewetting after the very dry months from October to February. Such observations about the effects of drying-rewetting on nitrate production have been frequently reported under laboratory (Fierer and Schimel 2002; Mian et al. 2008) as well as field conditions (Ford et al. 2007; Leitner et al. 2017).

During April, the rate of production of ammonium-N was low at both depths. It could be due to high temperature and very low soil moisture associated with very low rainfall and no irrigation; most of the ammonium-N is immobilized or volatilized (Whitehead and Raistrick 1991; Dempsey et al. 2017). Nitrate concentration was also low at both 0–10 and 10–20 cm depths in both pre- and post-incubated samples, so nitrification is not the cause. Because it is in the active growing season, a plausible reason could be that most of the available nitrate-N is taken up by plants, but lack of nitrate even after 7 days incubation for soils from both depths supports the idea of a sharp fall in nitrification in April as ammonium-N supply becomes restricted. The low production of ammonium-N in incubated soils in May and June but high levels of nitrate in soil pre-incubation in May suggest that inputs of ammonia from the atmosphere are being nitrified rather than ammonium being produced by mineralization of organic-N in the soil itself (Goulding et al. 1998).

During the hot months of May to July, ammonium-N decreases at both depths. No increase during incubation and loss of both ammonium-N and nitrate-N at both depths during incubation suggest volatilization, aided by an extremely low moisture level, because of the high temperature which limits most microbial activity. From November to December and on to January and February, the increases in the levels of soil organic C and N at both depths suggest the accumulation of litter in soil over winter leading to the increase in mineral-N production in March (Satti et al. 2003; Contosa et al. 2011). The concentrations of both ammonium-N and nitrate-N at both depths show why the idea of winter leaching is plausible because of good correlations between ammonium-N at 0–10 and 10–20 cm and between nitrate-N at 0–10 and 10–20 cm. Incubated soils show decreases in ammonium-N and nitrate-N at both depths, indicating immobilization and/or volatilized loss of ammonium-N. Clearly, over winter, atmospheric N inputs must be playing an important role as the N inputs, via atmospheric deposition, are generally higher during winter months (Wang et al. 2015).

Total mineral-N for both 0–10 and 10–20 cm soil depths showed winter peaks in immobilization/volatilization loss and relatively lower losses or small mineral-N inputs from soil during the summer season; both nitrate-N and ammonium-N contribute to defining the trends. However, ammonium-N contribution was greater which is why it shows good correlation with mineral-N concentrations both before and after incubation. Throughout the year, immobilization or loss of mineral-N was dominant and net mineralization inputs were very small during incubation. Both depths show similar trends except for September. Mineralization requires favorable moisture and temperature conditions that may only be occurring intermittently (Tian et al. 2010; Guntinas et al. 2012; Li et al. 2014).

The present research findings and previous studies suggested that N transformations are the net result of mineralization, denitrification, volatilization, and immobilization (Lovett and Rueth 1999; Singh et al. 2005). In a broader sense and wider context, these factors are dependent upon, and are interrelated to, temperature, humidity, and microbial activity (Dorana et al. 1998; Homyak et al. 2017; Joshi et al. 2017). Maximum amounts of nitrate-N and ammonium-N were found at transition periods, i.e., in autumn and spring due to change in environmental conditions. Seasonal variations are evident from the present study and in line with previous findings reporting strong seasonal N dynamics due to changes in climatic factors and anthropogenic N inputs (e.g., Corre et al. 2002; Zhang et al. 2008; Chen et al. 2017). Our study strongly supports the concept that the additions in N concentrations from soil by mineralization are often small and the N inputs from atmospheric deposition must be very important in regulating soil N cycling.

Conclusions

Despite the lack of very low temperatures at the Pakistan sites compared to those at York in the UK, there are some striking similarities in the results reported in this study. The site in Pakistan also gave significant strong maxima in ammonium production in March, which appears to correlate with increase in soil C and N contents over winter as litter accumulates, in spite of the higher winter temperatures compared to those in York in the UK. However, the flush of mineralization in March in Pakistan is short lived. In Pakistan, the dry conditions probably play a limiting role in winter months as much as the lower temperature. High plant N uptake and falling rates of production of ammonium and nitrate in soil in April as readily mineralizable organic matter from the previous winter becomes exhausted lead to very low soil nitrate concentration in April. The biggest difference between the UK and Pakistan sites is the greater relative importance of ammonium compared to nitrate in Pakistan. The soils at Faisalabad show statistically significant lower mineral-N concentrations in pre- and post-incubation in May and/or June compared with earlier months, following the March maxima. During incubation in summer months, the soils are immobilizing mineral-N rather than producing which is not what might have been intuitively expected under warm and moderately moist conditions. Future studies should take into account the role of atmospheric N deposition on soil N transformations while formulating the soil N and fertilization management strategies because N pulses from atmospheric deposition can have significant effects on soil N cycling in inherently nutrient-poor soils.