- Email: [email protected]
Ammonia loss following surface application of urea fertilizer to undrained and drained forested minerotrophic peatland sites in central Alberta, Canada
Pores;f;ology Management ELSEVIER
Forest
Ecology and Management 78 (1995) 139-145
Ammonia loss following surface application of urea fertilizer to undrained and drained forested minerotrophic peatland sites in central Alberta, Canada Aucelm G. Mugasha ap* , Donald J. Pluth b a Department
of Forest Biology, Faculty of Forestry, Sokoine University of Agriculture, P.O. Box 3010, Chuo Kikuu, b Department of Soil Science, University ofAlberta, Edmonton, Alta. T6G 2E3, Canada
Morogoro,
Tanzania
Accepted 16 May 1995
Abstract Drainage and urea fertilization of forested peatlands in western Canada are being contemplated as means for increasing peatland conifer productivity. Surface application of urea-based fertilizers may result in loss of some N by volatilization of NH,. The objective of the present study was to assess the loss of NH, from undrained and drained peatland sites as affected by various N dosages, i.e. 0 (unfertilized), 200 (N,) and 400 kg N ha-’ (NJ, following field broadcast application of urea fertilizer. Losses of NH, from fertilized plots were higher on the undrained relative to the drained site. Increase in urea-N dosage increased NH, volatilization from both sites. Most of the NH, volatilization losses occurred during the first 9 days after fertilizer application. The precipitation that occurred on day 9 after fertilizer application decreased NH, volatilization to background levels of unfertilized plots for the rest of the study period. Net mean losses of NH, from the undrained site were 3% and 4% of applied N for N, and N,, respectively. Corresponding losses on the drained site were 0.7% and 7%. These magnitudes of loss of N appear to be less important from a forest management point of view. Keywords:
Peatland;
Drainage;
Fertilization;
Urea; Ammonia
volatilization
1. Introduction
The growth of tamarack (Larix laricina (Du Roi) K. Koch.) and black spruce (Piceu mariana (Mill.) BSP.) in natural peatlands in western Canada is limited more often by nitrogen than by any other nutrient element (Lieffers and Macdonald, 1990; Mugasha et al., 1991, 1993). Nitrogen fertilization of both undrained and drained minerotrophic peatland sites increased foliar N concentration and unit needle
* Corresponding
author.
037%1127/95/$09.50 0 1995 Elsevier SSDI 0378-1127(95)03585-O
Science B.V. All rights reserved
mass of tamarack and black spruce (Mugasha et al., 1993). This suggests that N fertilization of peatlands may be an essential tool for improving conifer productivity. Urea-based fertilizers are commonly used for forest fertilization because they are highly concentrated, relatively less leaching occurs compared with nitrate fertilizer forms, and they are less toxic to man and animals than nitrate-containing N fertilizers (Nommik, 1973a,b). When urea fertilizer is added to forest soil, it is usually hydrolyzed to ammonium carbonate by soil urease. The rapid accumulation of ammonium and NH, and a corresponding rise in pH lead to gaseous loss of nitrogen as NH, (Marshall
A.G. Mugasha,
140
D.J. P&h/Forest
Ecology
and Management
N%(s) where
H+ + NH,(s)
(1)
NH&)
(2)
A
139-145
about pH 7, about 1% of NH,/NH: in solution is present as NH,, but the percentage increases as the pH is increased, as shown in Eq. (1). The partial pressure of NH, gas depends on the concentration of NH, in solution; it increases with increase of concentration of NH: ions, pH and temperature. Under field conditions, where the soil surface is open to the atmosphere, volatilization varies with the rate of transport of NH, away from the surface, which is determined by wind speed, and with evaporation of water from the surface (Wild, 1987). There is very little information on the extent of NH, volatilization from surface-applied urea-based fertilizers under peatland conditions. The present study was undertaken to estimate gaseous loss of applied N through NH, volatilization from a forest floor under tamarack/black spruce stand on contiguous undrained and drained minerotrophic peatland sites. The specific objective was to evaluate the degree of NH, volatilization from undrained and drained peatland sites as affected by urea-N dosages.
and DeBell, 1980). Nitrogen loss through volatilization from surface applied urea-based fertilizers is a potential problem affecting the efficiency of fertilizer urea in forestry. Estimates of losses from forest soils range from less than 5% (Overrein, 1968; Volk, 1970) to over 40% (Watkins et al., 1972; Nommik, 1973a). The factors involved in the process of NH, volatilization appear to be numerous (Macrae and Ancajas, 1970) and vary from place to place owing to variations in climate, vegetation, soil and experimental techniques (Watkins et al., 1972). Macrae and Ancajas (1970) noted that these factors are: (a) high soil pH; (b) high temperature; (c) high moisture loss; (d) low cation exchange capacity of the soil; (e) high organic matter content. Ammonia is lost under conditions which can be predicted from the following equations (Wild, 1988). NH; a
78 (1995)
[NH,] [H+1 Ka=
[NH:]
2. Materials
and methods
and the Henry coefficient, K, is The experimental area was part of the Wolf Creek Peatland Drainage Project located near Edson in central Alberta, Canada (53”25.6’N, 116”Ol’W). The area is within the Lower Boreal Cordiheran ecoregion (Corns and Annas, 1986) and is classified as a
p[NH,lg KH=
[NH,]S
p being the partial pressure of the gas, [NH,],, and [NH,], the concentration of NH, in solution. At
Table 1 Means of selected Site
soil properties
of undrained
Soil depth
Bulk density
(cm)
(Mg me31
O-10 10-20 20-30
0.020 (0.001) 0.044 (0.002) 0.066 (0.003)
O-10 10-20 20-30
0.019 (0.001) 0.049 (0.002) 0.086 (0.002)
and drained
sites before
fertilization
at Wolf Creek,
central Alberta
Electr. conduct. (dS m-‘)
Total
Total N (%o)
5.4 6.1 5.7
0.281 0.166 0.145
41.9 (0.29) 40.4 (0.45) 42.3 (0.25)
1.08 (0.034) 1.44 tO.096) 1.94 (0.080)
0.097 (0.004) 0.125 (0.004) 0.143 (0.004)
5.4 6.1 5.7
0.311 0.163 0.118
42.2 (0.23) 41.9 (0.23) 43.lCO.31)
1.03 (0.017) 1.77 (0.058) 2.55 (0.056)
0.093 (0.002) 0.138 (0.005) 0.144 (0.004)
a
Organic
Undrained
Drained
a Numbers of observations (n) per depth increment variables; standard errors are in parentheses. Source: Mugasha et al. (1993).
for bulk density
on undrained
sites were
respectively
60 and 108; n = 12 for other
A.G. Mugasha,
D.J. Pluth /Forest
Ecology
minerotrophic peatland or an intermediate fen based on water chemistry (Sjors, 1952). The Terric Fibric Mesisol (Agriculture Canada Expert Committee on Soil Survey, 1987) developed in 122 + 10 (SD) cm of peat. The mean stand density of overstory vegetation, i.e. tamarack (70 years) and black spruce (80 years) ranged from 1740 to 2240 stems ha-’ (basal area 3.6-4.5 m* ha-‘). The understory vegetation consisted of shrubs, herbs and mosses. Selected soil properties for drained and undrained sites are presented in Table 1. Two experiments were established, one on the undrained site and the other on the drained site. For each experiment, N fertilization consisted of the following levels: 0 kg N ha-’ (unfertilized control), 200 kg N ha-’ (N,) and 400 kg N ha-’ (NJ. The undrained site was upslope 150 m from the nearest drainage ditch of the drained site. The elevational difference between drained and undrained sites was about 1.5 m. Both sites were selected to minimize differences in stand density, understorey species composition and tree size. For the drained site, the ditch network was excavated with a bachhoe (Lannen S-10 with a contoured bucket) in fall 1987. The secondary drainage ditches were 40 m apart and about 90 cm deep. For each experiment, i.e. on each site, nine plots (three fertilization treatments X three replications), each 3 m X 3 m in size and at least 4 m apart were randomly laid out. On each site these plots were located at least 30 m away from the larger fertilization experiment described by Mugasha et al. (1993). On the drained site, the nine plots were located between 6 and 12 m from the secondary drainage ditch. Ammonia volatilization was measured by the microplot procedure using the modified semi-open NH, sorber technique described by Marshall and DeBell (1980). Ammonia volatilization from soil surface (moss surface) was monitored at nine alternate locations (microplots) within each 3 m X 3 m plot (as described below). Each microplot was made up of an open ended steel cylinder of 20 cm i.d. (base) and 30 cm height and a sorber chamber constructed from plastic pipe (20 cm i.d., 15 cm tall). The internal diameter of the base defined the area over which NH, was collected. The sorber chamber housed two horizontal polyfoam sorber disks (20 cm diameter, 2 cm thick) held in place by two 5 mm diameter
and Management
78 (1995)
139-145
141
aluminum pins driven through the center of the sorber chamber at 5 and 10 cm from the bottom. Ammonia evolved from the soil was captured by the lower sorber, ambient NH, by the upper sorber. Such an arrangement permitted interchange of other gases between the chamber and the surrounding atmosphere. The sorber was sheltered from throughfall by an inverted, disposable aluminum pie plate (Nason et al., 1988). Contamination of disks was prevented by adopting the procedure suggested by Marshall and DeBell (1980). Steel cylinders were installed in the intermediate micro-relief by cutting into the peat during summer 1988. The cylinders protruded 2-3 cm above the surface of live moss to allow the fitting of sorber chambers. Before fertilization all herbaceous plants within each microplot were removed by cutting at the surface of live moss. Fertilizer was applied to each plot (microplots inclusive) by broadcasting on 30 May 1989. For each plot (i.e. for each fertilization treatment replicate) on each site, three sorber chambers were fitted to three steel cylinders (first set of the three sets per plot) and secured in place by rubber bands (3 cm wide). The sorber disks (polyfoams) were charged by soaking in 0.75 M phosphoric acid in 2.5% glycerol and allowing the excess solution to drain under gravity (Nason et al., 1988). Sorber disks in each sorber chamber were replaced after 48 or 72 h for a period of 26 days after fertilizer application. After the first replacement of sorber disks, chambers were transferred to a second set of three microplots within each plot. After the second replacement of sorber disks, sorber chambers were transferred to a third set of microplots within each plot. This procedure was repeated until the experiment was terminated. All retrieved sorber disks were stored at - 20°C for future chemical analysis. In the laboratory, disk sorbers were thawed and, for the extraction of sorbed NH,, were rinsed and squeezed dry with 4 X 100 ml of deionized water. Extracts were poured into the volumetric flask and made up to volume with deionized water. NH:-N in the extracts was determined by automated analysis (Technicon Instruments, 1977). Statistical analyses were carried out using the general linear models (GLM) procedure of the Statistical Analysis Systems Institute Inc. @AS Institute Inc., 1987). For each plot, NH, evolved from each
142
A.G. Mugasha,
D.J. Pluth / Forest
Ecology
and Management
78 (1995)
139-145
set of the three microplots was averaged for each sorber disk retrieval date (sampling date). Thereafter, periodic and cumulative NH, volatilization values (kg N ha-‘) were calculated per plot. For each sampling date, each data set was subjected to the following statistical analysis. For each experiment (i.e. each drainage site) and each sampling date, the ANOVA model tested for the effect of fertilization. Prior to analysis of variance, the null hypothesis for the homogeneity of each error variance was tested by the Shapiro-Wilk procedure (SAS Institute Inc., 1987). The Box and Cox (1964) procedure was used to select the appropriate transformation (natural logarithm) since the error variances were generally heterogenous.For significant treatmentsand within each drainage site, the differences between fertilizer levels and each sampling date were examined by Duncan’s multiple range test @AS Institute Inc., 1987).
3. Results and discussion The amounts of NH,-N volatilized and trapped from the three fertilization treatments on the undrained and drained sites during the study period are shown in Fig. 1. Although not tested statistically, lossesof NH, from unfertilized plots on undrained and drained sites were similar. The average amounts of NH, which volatilized from unfertilized-undrained and unfertilized-drained sites were 0.04 kg and 0.033 kg N ha- ’ day-‘, respectively. We are not aware of any other study that simultaneously evaluated NH, volatilization from drained and undrained peatland sites. However, our results are in sharp contrast with those reported from upland clay loam soils (Willis and Sturgis, 1944). They observed that NH, volatilization from flooded soil was much larger than that of unflooded soil. For both drained and undrained sites, ammonia lossesfrom fertilized plots differed significantly from that from the unfertilized control plots (Fig. 1, Table 2). Net cumulative NH, volatilization lossesduring the 26 days after N application on the undrained site were 5.92 kg (3% of applied N) and 16.1 kg N ha-’ (4%) for N, and N,, respectively. Corresponding losseson the drained site were 1.34 kg (0.7%) and 13.2 kg N ha-’ (3.3%). Although not tested statistically, ammonia losses
Days
after
fertilization
Fig. 1. Daily precipitation and mean daily air temperature (a). periodic (b) and cumulative (c) volatilization of ammonia from unfertilized and surface applied urea fertilizer on undrained and drained peatland sites at Wolf Creek, central Alberta, Canada. Urea-N fertilizer was applied by broadcasting at rates of 200 and 400 kg N ha-’ on 30 May 1989. Each point is the mean of three replications. In (a), hanging bars indicate daily precipitation events, and continuous lines indicate daily mean air temperature. The undrained site is indicated by closed symbols and the drained site by open symbols. Circular, square and triangular symbols represent the control (unfertilized), N, and N, fertilization treatments, respectively.
A.G. Mugasha,
D.J. Pluth/Forest
Ecology
and Management
Site a
Drained Undrained a Degrees
of periodic
Time after fertilization
volatilization
of ammonia
139-145
143
volatilization on drained-N, microplots was higher than that of the other treatments between 1 and 4 days after fertilizer application. This may be related to a greater increase in pH following ureolysis and higher temperature of the soil surface. On the drained site, mean soil temperature (l-2 cm below the moss surface) at 12:00 and 14:00 h within 4 days after fertilizer application was 4.6”C higher than that on the undrained site. Our results show that NH, volatilization was a rapid process that was completed between 6 and 9 days after N application. After this period there was no significant increase in periodic or cumulative NH, loss. Prior to day 9 there was no precipitation, then on this day there was 40.64 mm of rainfall. This may have washed NH, and ammonium into the soil on the drained site. However, on the undrained site, flooding of microplots by rainfall probably washed away some of the fertilizer. Periodic losses of ammonia from fertilized microplots show some slight correlation with daily air temperature and high correlation with precipitation (Figs. l(a) and l(b)). This could be due to the interaction of soil surface moisture and temperature. It is also possible that shading by sorber chamber during gaseous exchange measurements might have confounded the effect of ambient temperature and precipitation.
from fertilized plots were relatively higher on the undrained compared with the drained site (Fig. 1). We are not aware of any study that simultaneously examined NH, volatilization from fertilized flooded natural and drained peatland sites. However, some research on NH, loss has been carried out on flooded and dry mineral soils. Rajaratham and Purushothaman (1973) attributed the greater amount of NH, volatilized under waterlogged conditions to the greater evaporation of water, NH, being soluble in water. Wahhab et al. (19.57) found a positive correlation between water evaporation and NH, volatilized. In contrast, Chao and Kroontje (1964) concluded that NH, volatilization and water evaporation follow different functions. Under flooded conditions, Ventura and Yoshida (1977) attributed the increase in NH, volatilization to an increase in soil pH. They observed that the continually flooded soil had a higher pH than that of dry soil (i.e. pH 7.1 vs. 6.6). The amount of NH, volatilized increased with the increase in urea-N dosage (Fig. 1, Table 2). There was also a trend of increasing percentage of NH, that volatilized with higher dosage of urea. Our results, concerning the increase in NH, losses in relation to N dosage, are in agreement with those reported by Nelson (1982). However, Nelson’s results showed that the proportion of added N that is volatilized as NH, remains constant across the range of N application. The effects of fertilization on periodic NH, volatilization on each site was apparent (Fig. l(b)) from 2 to 9 days following fertilizer application. Thereafter, there was no substantial effect of fertilization on NH, volatilization. The periodic and cumulative NH, loss patterns on the undrained and drained fertilized sites were very similar, except for the N, treatment on the drained site. Ammonia
Table 2 Probability values from ANOVA Alberta, Canada
78 (1995)
4. Concluding
remarks
This experiment was conducted under field conditions in order to provide a preliminary estimate of potential NH, volatilization from surface applied urea fertilizer on undrained and drained peatland sites. The quantities of NH, volatilized in this study
in response
to drainage
and urea fertilization
at Wolf
Creek,
(days)
2
4
6
9
12
16
20
23
26
0.011 0.332
0.137 0.031
0.183 0.024
< 0.001 0.001
0.050 0.005
0.319 0.022
0.029 0.020
0.143 0.060
0.825 0.433
of freedom
for drainage,
fertilization
and drainage
X fertilization
interaction
were 1, 2 and 2, respectively.
Central
144
A.G. Mugasha,
D.J. Pluth /Forest
Ecology
could have been slightly underestimated. Marshall and DeBell (1980) observed that NH, volatilization estimated by the open system (‘5N-balance) was 20% higher than that estimated by the semi-open technique (used in this study). In spite of this limitation, some inferences can be made from our study. (1) Ammonia volatilization was a rapid process that was completed in 6-9 days after fertilizer application. (2) Ammonia volatilization from fertilized plots was higher on the undrained than the drained site. (3) Ammonia volatilization increased with increase in urea-N dosage. (4) The amounts of NH, volatilized averaged 3.5% and 4% of applied fertilizer on undrained and drained sites, respectively. A loss of this magnitude would be unimportant, especially in view of the possibility that urea will be retained by live moss and organic matter of surface soil.
Acknowledgements
The present work was done when the author was a graduate student at the University of Alberta. Financial support was provided by Natural Science and Engineering Research Council of Canada to Drs. V.J. Lieffers and R.L. Rothwell, Department of Forest Science, University of Alberta; the Canadian Forestry Service Human Resources Development Grant to and the Department of Soil Science, University of Alberta for research/ teaching assistantships to A.G.M.; and the Canada/Alberta Forestry Resources Development Agreement, Peatland Drainage and Improvement Program. We would like to thank K. Greenway, A. Maenpaa, and T. Rothwell for field assistance, and J. Konwicki for chemical analysis. We wish to express our sincere thanks and appreciation for this support.
References Agriculture Canada Expert Committee on Soil Survey, 1987. The Canadian System of Soil Classification, 2nd edn. Publ. No. 1646, Agriculture Canada, Ottawa.
and Management
78 (1995)
139-145
Box, G.E.P. and Cox, D.R., 1964. An analysis of transformations. J. R. Stat. Sot., Ser. B, 26: 211-243. Chao, T. and Kroontje, W., 1964. Relationship between ammonia volatilization, ammonia concentration and water evaporation. Soil Sci. Sot. Am. Proc., 28: 393-395. Corns, I.G.W. and Annas, R.M., 1986. Field guide to forest ecosystems of west-central Alberta. Canadian Forest Service, Northern Forestry Center, Edmonton, Alta. Lieffers, V.J. and Macdonald, ES., 1990. Growth and foliar nutrient status of black spruce and tamarack in relation to depth to water table in some Alberta peatlands. Can. J. For Res., 20: 805-809. Macrae, I.C. and Ancajas, R., 1970. Volatilization of ammonia from submerged tropical soils. Plant Soil, 33: 97-103. Marshall, V.G. and DeBelI, D.S., 1980. Comparison of four methods of measuring volatilization losses of nitrogen following urea fertilization of forest soils. Can. J. Soil Sci., 60: 549-563. Mugasha, A.G., Pluth, D.J., Higginbotbam, K.O. and Takyi, SK., 1991. Foliar responses of blackspruce to thinning and fertilization of a drained peatland. Can. J. For. Res., 21: 152-163. Mugasha. A.G., Pluth, D.J. and Hillman, G.R., 1993. Foliar responses of tamarack and black spruce to drainage and fertilization of a minerotrophic peatland. Can. J. For. Res., 23: 166-180. Nason, G.E., Pluth, D.J. and McGill, W.B., 1988. Volatilization and foliar recapture of ammonia following spring and fall application of nitrogen-15 urea to a douglas-fir ecosystem. Soil Sci. Sot. Am. J., 52: 821-828. Nelson, D.W., 1982. Gaseous losses of nitrogen other than through denitrification. In: F.J. Steven (Editor), Nitrogen in Agricultural Soils. Agronoour, 22: 327-364. Nommik, H., 1973a. The effect of pellet on the ammonia loss from urea applied to forest soil. Plant Soil, 39: 309-318. Nommik, H., 1973b. Assessment of volatilization loss of ammonia from surface applied urea on forest soil by 15N recovery. Plant Soil, 38: 589-603. Overrein, L.N., 1968. Lysimeter studies on tracer nitrogen in forest soils: I: Nitrogen losses by leaching and volatilizatioe after addition of urea-, ammoniumand nitrate-” N. Soil Sci,, 107: 149-159. Rajaratham, J.A. and Purushothaman, V., 1973. Studies of nitrogen losses from soils in Malaysia. I. Influence of soil moisture, rates and types of nitrogenous compound on ammonia volatilization. Malay. Agric. Res., 2: 59-64. Sjors, H., 1952. On the relations between vegetation and electrolytes in north Swedish mire waters. Oikos, 2: 241-258. Statistical Analysis Systems Institute Inc., 1987. Stat guide to personal computers, 6th edn. SAS Institute Inc., Cary, NC. Technicon Instruments, 1977. Ammonia in water and wastewater. Industrial method No. 98-7OW, Technicon Industrial Systems/ Tarrytown, NY. Ventura, W.B. and Yoshida, Y., 1977. Ammonia volatilization from a flooded tropical soil. Plant Soil, 46: 521-531. Volk, G.M., 1970. Gaseous loss of ammonia from prilled urea applied to slash pine. Soil Sci. Sot. Am. Proc., 34: 513-516.
A.G. Mugasha,
DJ. Pluth/Forest
Ecology
Wahhab, A, Randhawa, M.S. and Alam, S.Q., 1957. Losses of ammonia from ammonium sulphate under different conditions when applied to soil. Soil Sci., 84: 249-255. Watkins, S.H., Strand, R.F., DeBell, D.S. and Esch, J., 1972. Factors infhrencing ammonia losses from urea applied to northwestern forest soils. Soil Sci. Sot. Am. J., 36: 353-357.
and Management
78 (1995)
139-145
145
Wild, A., 1987. Plant nutrients in soil: nitrogen. In: A. Wild (Editor), Russel’s Soil Conditions and Plant Growth, 7th edn. longman/Wiley, New York, pp. 652-694. Willis, W.H. and Sturgis, M.B., 1944. Loss of nitrogen from soil affected by changes in temperatures and reaction. Soil Sci. Sot. Amer. Proc., 9: 106-113.
Forest
Ecology and Management 78 (1995) 139-145
Ammonia loss following surface application of urea fertilizer to undrained and drained forested minerotrophic peatland sites in central Alberta, Canada Aucelm G. Mugasha ap* , Donald J. Pluth b a Department
of Forest Biology, Faculty of Forestry, Sokoine University of Agriculture, P.O. Box 3010, Chuo Kikuu, b Department of Soil Science, University ofAlberta, Edmonton, Alta. T6G 2E3, Canada
Morogoro,
Tanzania
Accepted 16 May 1995
Abstract Drainage and urea fertilization of forested peatlands in western Canada are being contemplated as means for increasing peatland conifer productivity. Surface application of urea-based fertilizers may result in loss of some N by volatilization of NH,. The objective of the present study was to assess the loss of NH, from undrained and drained peatland sites as affected by various N dosages, i.e. 0 (unfertilized), 200 (N,) and 400 kg N ha-’ (NJ, following field broadcast application of urea fertilizer. Losses of NH, from fertilized plots were higher on the undrained relative to the drained site. Increase in urea-N dosage increased NH, volatilization from both sites. Most of the NH, volatilization losses occurred during the first 9 days after fertilizer application. The precipitation that occurred on day 9 after fertilizer application decreased NH, volatilization to background levels of unfertilized plots for the rest of the study period. Net mean losses of NH, from the undrained site were 3% and 4% of applied N for N, and N,, respectively. Corresponding losses on the drained site were 0.7% and 7%. These magnitudes of loss of N appear to be less important from a forest management point of view. Keywords:
Peatland;
Drainage;
Fertilization;
Urea; Ammonia
volatilization
1. Introduction
The growth of tamarack (Larix laricina (Du Roi) K. Koch.) and black spruce (Piceu mariana (Mill.) BSP.) in natural peatlands in western Canada is limited more often by nitrogen than by any other nutrient element (Lieffers and Macdonald, 1990; Mugasha et al., 1991, 1993). Nitrogen fertilization of both undrained and drained minerotrophic peatland sites increased foliar N concentration and unit needle
* Corresponding
author.
037%1127/95/$09.50 0 1995 Elsevier SSDI 0378-1127(95)03585-O
Science B.V. All rights reserved
mass of tamarack and black spruce (Mugasha et al., 1993). This suggests that N fertilization of peatlands may be an essential tool for improving conifer productivity. Urea-based fertilizers are commonly used for forest fertilization because they are highly concentrated, relatively less leaching occurs compared with nitrate fertilizer forms, and they are less toxic to man and animals than nitrate-containing N fertilizers (Nommik, 1973a,b). When urea fertilizer is added to forest soil, it is usually hydrolyzed to ammonium carbonate by soil urease. The rapid accumulation of ammonium and NH, and a corresponding rise in pH lead to gaseous loss of nitrogen as NH, (Marshall
A.G. Mugasha,
140
D.J. P&h/Forest
Ecology
and Management
N%(s) where
H+ + NH,(s)
(1)
NH&)
(2)
A
139-145
about pH 7, about 1% of NH,/NH: in solution is present as NH,, but the percentage increases as the pH is increased, as shown in Eq. (1). The partial pressure of NH, gas depends on the concentration of NH, in solution; it increases with increase of concentration of NH: ions, pH and temperature. Under field conditions, where the soil surface is open to the atmosphere, volatilization varies with the rate of transport of NH, away from the surface, which is determined by wind speed, and with evaporation of water from the surface (Wild, 1987). There is very little information on the extent of NH, volatilization from surface-applied urea-based fertilizers under peatland conditions. The present study was undertaken to estimate gaseous loss of applied N through NH, volatilization from a forest floor under tamarack/black spruce stand on contiguous undrained and drained minerotrophic peatland sites. The specific objective was to evaluate the degree of NH, volatilization from undrained and drained peatland sites as affected by urea-N dosages.
and DeBell, 1980). Nitrogen loss through volatilization from surface applied urea-based fertilizers is a potential problem affecting the efficiency of fertilizer urea in forestry. Estimates of losses from forest soils range from less than 5% (Overrein, 1968; Volk, 1970) to over 40% (Watkins et al., 1972; Nommik, 1973a). The factors involved in the process of NH, volatilization appear to be numerous (Macrae and Ancajas, 1970) and vary from place to place owing to variations in climate, vegetation, soil and experimental techniques (Watkins et al., 1972). Macrae and Ancajas (1970) noted that these factors are: (a) high soil pH; (b) high temperature; (c) high moisture loss; (d) low cation exchange capacity of the soil; (e) high organic matter content. Ammonia is lost under conditions which can be predicted from the following equations (Wild, 1988). NH; a
78 (1995)
[NH,] [H+1 Ka=
[NH:]
2. Materials
and methods
and the Henry coefficient, K, is The experimental area was part of the Wolf Creek Peatland Drainage Project located near Edson in central Alberta, Canada (53”25.6’N, 116”Ol’W). The area is within the Lower Boreal Cordiheran ecoregion (Corns and Annas, 1986) and is classified as a
p[NH,lg KH=
[NH,]S
p being the partial pressure of the gas, [NH,],, and [NH,], the concentration of NH, in solution. At
Table 1 Means of selected Site
soil properties
of undrained
Soil depth
Bulk density
(cm)
(Mg me31
O-10 10-20 20-30
0.020 (0.001) 0.044 (0.002) 0.066 (0.003)
O-10 10-20 20-30
0.019 (0.001) 0.049 (0.002) 0.086 (0.002)
and drained
sites before
fertilization
at Wolf Creek,
central Alberta
Electr. conduct. (dS m-‘)
Total
Total N (%o)
5.4 6.1 5.7
0.281 0.166 0.145
41.9 (0.29) 40.4 (0.45) 42.3 (0.25)
1.08 (0.034) 1.44 tO.096) 1.94 (0.080)
0.097 (0.004) 0.125 (0.004) 0.143 (0.004)
5.4 6.1 5.7
0.311 0.163 0.118
42.2 (0.23) 41.9 (0.23) 43.lCO.31)
1.03 (0.017) 1.77 (0.058) 2.55 (0.056)
0.093 (0.002) 0.138 (0.005) 0.144 (0.004)
a
Organic
Undrained
Drained
a Numbers of observations (n) per depth increment variables; standard errors are in parentheses. Source: Mugasha et al. (1993).
for bulk density
on undrained
sites were
respectively
60 and 108; n = 12 for other
A.G. Mugasha,
D.J. Pluth /Forest
Ecology
minerotrophic peatland or an intermediate fen based on water chemistry (Sjors, 1952). The Terric Fibric Mesisol (Agriculture Canada Expert Committee on Soil Survey, 1987) developed in 122 + 10 (SD) cm of peat. The mean stand density of overstory vegetation, i.e. tamarack (70 years) and black spruce (80 years) ranged from 1740 to 2240 stems ha-’ (basal area 3.6-4.5 m* ha-‘). The understory vegetation consisted of shrubs, herbs and mosses. Selected soil properties for drained and undrained sites are presented in Table 1. Two experiments were established, one on the undrained site and the other on the drained site. For each experiment, N fertilization consisted of the following levels: 0 kg N ha-’ (unfertilized control), 200 kg N ha-’ (N,) and 400 kg N ha-’ (NJ. The undrained site was upslope 150 m from the nearest drainage ditch of the drained site. The elevational difference between drained and undrained sites was about 1.5 m. Both sites were selected to minimize differences in stand density, understorey species composition and tree size. For the drained site, the ditch network was excavated with a bachhoe (Lannen S-10 with a contoured bucket) in fall 1987. The secondary drainage ditches were 40 m apart and about 90 cm deep. For each experiment, i.e. on each site, nine plots (three fertilization treatments X three replications), each 3 m X 3 m in size and at least 4 m apart were randomly laid out. On each site these plots were located at least 30 m away from the larger fertilization experiment described by Mugasha et al. (1993). On the drained site, the nine plots were located between 6 and 12 m from the secondary drainage ditch. Ammonia volatilization was measured by the microplot procedure using the modified semi-open NH, sorber technique described by Marshall and DeBell (1980). Ammonia volatilization from soil surface (moss surface) was monitored at nine alternate locations (microplots) within each 3 m X 3 m plot (as described below). Each microplot was made up of an open ended steel cylinder of 20 cm i.d. (base) and 30 cm height and a sorber chamber constructed from plastic pipe (20 cm i.d., 15 cm tall). The internal diameter of the base defined the area over which NH, was collected. The sorber chamber housed two horizontal polyfoam sorber disks (20 cm diameter, 2 cm thick) held in place by two 5 mm diameter
and Management
78 (1995)
139-145
141
aluminum pins driven through the center of the sorber chamber at 5 and 10 cm from the bottom. Ammonia evolved from the soil was captured by the lower sorber, ambient NH, by the upper sorber. Such an arrangement permitted interchange of other gases between the chamber and the surrounding atmosphere. The sorber was sheltered from throughfall by an inverted, disposable aluminum pie plate (Nason et al., 1988). Contamination of disks was prevented by adopting the procedure suggested by Marshall and DeBell (1980). Steel cylinders were installed in the intermediate micro-relief by cutting into the peat during summer 1988. The cylinders protruded 2-3 cm above the surface of live moss to allow the fitting of sorber chambers. Before fertilization all herbaceous plants within each microplot were removed by cutting at the surface of live moss. Fertilizer was applied to each plot (microplots inclusive) by broadcasting on 30 May 1989. For each plot (i.e. for each fertilization treatment replicate) on each site, three sorber chambers were fitted to three steel cylinders (first set of the three sets per plot) and secured in place by rubber bands (3 cm wide). The sorber disks (polyfoams) were charged by soaking in 0.75 M phosphoric acid in 2.5% glycerol and allowing the excess solution to drain under gravity (Nason et al., 1988). Sorber disks in each sorber chamber were replaced after 48 or 72 h for a period of 26 days after fertilizer application. After the first replacement of sorber disks, chambers were transferred to a second set of three microplots within each plot. After the second replacement of sorber disks, sorber chambers were transferred to a third set of microplots within each plot. This procedure was repeated until the experiment was terminated. All retrieved sorber disks were stored at - 20°C for future chemical analysis. In the laboratory, disk sorbers were thawed and, for the extraction of sorbed NH,, were rinsed and squeezed dry with 4 X 100 ml of deionized water. Extracts were poured into the volumetric flask and made up to volume with deionized water. NH:-N in the extracts was determined by automated analysis (Technicon Instruments, 1977). Statistical analyses were carried out using the general linear models (GLM) procedure of the Statistical Analysis Systems Institute Inc. @AS Institute Inc., 1987). For each plot, NH, evolved from each
142
A.G. Mugasha,
D.J. Pluth / Forest
Ecology
and Management
78 (1995)
139-145
set of the three microplots was averaged for each sorber disk retrieval date (sampling date). Thereafter, periodic and cumulative NH, volatilization values (kg N ha-‘) were calculated per plot. For each sampling date, each data set was subjected to the following statistical analysis. For each experiment (i.e. each drainage site) and each sampling date, the ANOVA model tested for the effect of fertilization. Prior to analysis of variance, the null hypothesis for the homogeneity of each error variance was tested by the Shapiro-Wilk procedure (SAS Institute Inc., 1987). The Box and Cox (1964) procedure was used to select the appropriate transformation (natural logarithm) since the error variances were generally heterogenous.For significant treatmentsand within each drainage site, the differences between fertilizer levels and each sampling date were examined by Duncan’s multiple range test @AS Institute Inc., 1987).
3. Results and discussion The amounts of NH,-N volatilized and trapped from the three fertilization treatments on the undrained and drained sites during the study period are shown in Fig. 1. Although not tested statistically, lossesof NH, from unfertilized plots on undrained and drained sites were similar. The average amounts of NH, which volatilized from unfertilized-undrained and unfertilized-drained sites were 0.04 kg and 0.033 kg N ha- ’ day-‘, respectively. We are not aware of any other study that simultaneously evaluated NH, volatilization from drained and undrained peatland sites. However, our results are in sharp contrast with those reported from upland clay loam soils (Willis and Sturgis, 1944). They observed that NH, volatilization from flooded soil was much larger than that of unflooded soil. For both drained and undrained sites, ammonia lossesfrom fertilized plots differed significantly from that from the unfertilized control plots (Fig. 1, Table 2). Net cumulative NH, volatilization lossesduring the 26 days after N application on the undrained site were 5.92 kg (3% of applied N) and 16.1 kg N ha-’ (4%) for N, and N,, respectively. Corresponding losseson the drained site were 1.34 kg (0.7%) and 13.2 kg N ha-’ (3.3%). Although not tested statistically, ammonia losses
Days
after
fertilization
Fig. 1. Daily precipitation and mean daily air temperature (a). periodic (b) and cumulative (c) volatilization of ammonia from unfertilized and surface applied urea fertilizer on undrained and drained peatland sites at Wolf Creek, central Alberta, Canada. Urea-N fertilizer was applied by broadcasting at rates of 200 and 400 kg N ha-’ on 30 May 1989. Each point is the mean of three replications. In (a), hanging bars indicate daily precipitation events, and continuous lines indicate daily mean air temperature. The undrained site is indicated by closed symbols and the drained site by open symbols. Circular, square and triangular symbols represent the control (unfertilized), N, and N, fertilization treatments, respectively.
A.G. Mugasha,
D.J. Pluth/Forest
Ecology
and Management
Site a
Drained Undrained a Degrees
of periodic
Time after fertilization
volatilization
of ammonia
139-145
143
volatilization on drained-N, microplots was higher than that of the other treatments between 1 and 4 days after fertilizer application. This may be related to a greater increase in pH following ureolysis and higher temperature of the soil surface. On the drained site, mean soil temperature (l-2 cm below the moss surface) at 12:00 and 14:00 h within 4 days after fertilizer application was 4.6”C higher than that on the undrained site. Our results show that NH, volatilization was a rapid process that was completed between 6 and 9 days after N application. After this period there was no significant increase in periodic or cumulative NH, loss. Prior to day 9 there was no precipitation, then on this day there was 40.64 mm of rainfall. This may have washed NH, and ammonium into the soil on the drained site. However, on the undrained site, flooding of microplots by rainfall probably washed away some of the fertilizer. Periodic losses of ammonia from fertilized microplots show some slight correlation with daily air temperature and high correlation with precipitation (Figs. l(a) and l(b)). This could be due to the interaction of soil surface moisture and temperature. It is also possible that shading by sorber chamber during gaseous exchange measurements might have confounded the effect of ambient temperature and precipitation.
from fertilized plots were relatively higher on the undrained compared with the drained site (Fig. 1). We are not aware of any study that simultaneously examined NH, volatilization from fertilized flooded natural and drained peatland sites. However, some research on NH, loss has been carried out on flooded and dry mineral soils. Rajaratham and Purushothaman (1973) attributed the greater amount of NH, volatilized under waterlogged conditions to the greater evaporation of water, NH, being soluble in water. Wahhab et al. (19.57) found a positive correlation between water evaporation and NH, volatilized. In contrast, Chao and Kroontje (1964) concluded that NH, volatilization and water evaporation follow different functions. Under flooded conditions, Ventura and Yoshida (1977) attributed the increase in NH, volatilization to an increase in soil pH. They observed that the continually flooded soil had a higher pH than that of dry soil (i.e. pH 7.1 vs. 6.6). The amount of NH, volatilized increased with the increase in urea-N dosage (Fig. 1, Table 2). There was also a trend of increasing percentage of NH, that volatilized with higher dosage of urea. Our results, concerning the increase in NH, losses in relation to N dosage, are in agreement with those reported by Nelson (1982). However, Nelson’s results showed that the proportion of added N that is volatilized as NH, remains constant across the range of N application. The effects of fertilization on periodic NH, volatilization on each site was apparent (Fig. l(b)) from 2 to 9 days following fertilizer application. Thereafter, there was no substantial effect of fertilization on NH, volatilization. The periodic and cumulative NH, loss patterns on the undrained and drained fertilized sites were very similar, except for the N, treatment on the drained site. Ammonia
Table 2 Probability values from ANOVA Alberta, Canada
78 (1995)
4. Concluding
remarks
This experiment was conducted under field conditions in order to provide a preliminary estimate of potential NH, volatilization from surface applied urea fertilizer on undrained and drained peatland sites. The quantities of NH, volatilized in this study
in response
to drainage
and urea fertilization
at Wolf
Creek,
(days)
2
4
6
9
12
16
20
23
26
0.011 0.332
0.137 0.031
0.183 0.024
< 0.001 0.001
0.050 0.005
0.319 0.022
0.029 0.020
0.143 0.060
0.825 0.433
of freedom
for drainage,
fertilization
and drainage
X fertilization
interaction
were 1, 2 and 2, respectively.
Central
144
A.G. Mugasha,
D.J. Pluth /Forest
Ecology
could have been slightly underestimated. Marshall and DeBell (1980) observed that NH, volatilization estimated by the open system (‘5N-balance) was 20% higher than that estimated by the semi-open technique (used in this study). In spite of this limitation, some inferences can be made from our study. (1) Ammonia volatilization was a rapid process that was completed in 6-9 days after fertilizer application. (2) Ammonia volatilization from fertilized plots was higher on the undrained than the drained site. (3) Ammonia volatilization increased with increase in urea-N dosage. (4) The amounts of NH, volatilized averaged 3.5% and 4% of applied fertilizer on undrained and drained sites, respectively. A loss of this magnitude would be unimportant, especially in view of the possibility that urea will be retained by live moss and organic matter of surface soil.
Acknowledgements
The present work was done when the author was a graduate student at the University of Alberta. Financial support was provided by Natural Science and Engineering Research Council of Canada to Drs. V.J. Lieffers and R.L. Rothwell, Department of Forest Science, University of Alberta; the Canadian Forestry Service Human Resources Development Grant to and the Department of Soil Science, University of Alberta for research/ teaching assistantships to A.G.M.; and the Canada/Alberta Forestry Resources Development Agreement, Peatland Drainage and Improvement Program. We would like to thank K. Greenway, A. Maenpaa, and T. Rothwell for field assistance, and J. Konwicki for chemical analysis. We wish to express our sincere thanks and appreciation for this support.
References Agriculture Canada Expert Committee on Soil Survey, 1987. The Canadian System of Soil Classification, 2nd edn. Publ. No. 1646, Agriculture Canada, Ottawa.
and Management
78 (1995)
139-145
Box, G.E.P. and Cox, D.R., 1964. An analysis of transformations. J. R. Stat. Sot., Ser. B, 26: 211-243. Chao, T. and Kroontje, W., 1964. Relationship between ammonia volatilization, ammonia concentration and water evaporation. Soil Sci. Sot. Am. Proc., 28: 393-395. Corns, I.G.W. and Annas, R.M., 1986. Field guide to forest ecosystems of west-central Alberta. Canadian Forest Service, Northern Forestry Center, Edmonton, Alta. Lieffers, V.J. and Macdonald, ES., 1990. Growth and foliar nutrient status of black spruce and tamarack in relation to depth to water table in some Alberta peatlands. Can. J. For Res., 20: 805-809. Macrae, I.C. and Ancajas, R., 1970. Volatilization of ammonia from submerged tropical soils. Plant Soil, 33: 97-103. Marshall, V.G. and DeBelI, D.S., 1980. Comparison of four methods of measuring volatilization losses of nitrogen following urea fertilization of forest soils. Can. J. Soil Sci., 60: 549-563. Mugasha, A.G., Pluth, D.J., Higginbotbam, K.O. and Takyi, SK., 1991. Foliar responses of blackspruce to thinning and fertilization of a drained peatland. Can. J. For. Res., 21: 152-163. Mugasha. A.G., Pluth, D.J. and Hillman, G.R., 1993. Foliar responses of tamarack and black spruce to drainage and fertilization of a minerotrophic peatland. Can. J. For. Res., 23: 166-180. Nason, G.E., Pluth, D.J. and McGill, W.B., 1988. Volatilization and foliar recapture of ammonia following spring and fall application of nitrogen-15 urea to a douglas-fir ecosystem. Soil Sci. Sot. Am. J., 52: 821-828. Nelson, D.W., 1982. Gaseous losses of nitrogen other than through denitrification. In: F.J. Steven (Editor), Nitrogen in Agricultural Soils. Agronoour, 22: 327-364. Nommik, H., 1973a. The effect of pellet on the ammonia loss from urea applied to forest soil. Plant Soil, 39: 309-318. Nommik, H., 1973b. Assessment of volatilization loss of ammonia from surface applied urea on forest soil by 15N recovery. Plant Soil, 38: 589-603. Overrein, L.N., 1968. Lysimeter studies on tracer nitrogen in forest soils: I: Nitrogen losses by leaching and volatilizatioe after addition of urea-, ammoniumand nitrate-” N. Soil Sci,, 107: 149-159. Rajaratham, J.A. and Purushothaman, V., 1973. Studies of nitrogen losses from soils in Malaysia. I. Influence of soil moisture, rates and types of nitrogenous compound on ammonia volatilization. Malay. Agric. Res., 2: 59-64. Sjors, H., 1952. On the relations between vegetation and electrolytes in north Swedish mire waters. Oikos, 2: 241-258. Statistical Analysis Systems Institute Inc., 1987. Stat guide to personal computers, 6th edn. SAS Institute Inc., Cary, NC. Technicon Instruments, 1977. Ammonia in water and wastewater. Industrial method No. 98-7OW, Technicon Industrial Systems/ Tarrytown, NY. Ventura, W.B. and Yoshida, Y., 1977. Ammonia volatilization from a flooded tropical soil. Plant Soil, 46: 521-531. Volk, G.M., 1970. Gaseous loss of ammonia from prilled urea applied to slash pine. Soil Sci. Sot. Am. Proc., 34: 513-516.
A.G. Mugasha,
DJ. Pluth/Forest
Ecology
Wahhab, A, Randhawa, M.S. and Alam, S.Q., 1957. Losses of ammonia from ammonium sulphate under different conditions when applied to soil. Soil Sci., 84: 249-255. Watkins, S.H., Strand, R.F., DeBell, D.S. and Esch, J., 1972. Factors infhrencing ammonia losses from urea applied to northwestern forest soils. Soil Sci. Sot. Am. J., 36: 353-357.
and Management
78 (1995)
139-145
145
Wild, A., 1987. Plant nutrients in soil: nitrogen. In: A. Wild (Editor), Russel’s Soil Conditions and Plant Growth, 7th edn. longman/Wiley, New York, pp. 652-694. Willis, W.H. and Sturgis, M.B., 1944. Loss of nitrogen from soil affected by changes in temperatures and reaction. Soil Sci. Sot. Amer. Proc., 9: 106-113.