Tải bản đầy đủ - 0 (trang)
B. Nutrients in the Roots and Crop Residues

B. Nutrients in the Roots and Crop Residues

Tải bản đầy đủ - 0trang

ASSESSING SOIL FERTILITY DECLINE IN THE TROPICS



203



Table IX

Nutrient Uptake (kg h1 Ỉ 1 SD) of Sweet Potato at Two Sites in the

Humid Lowlands of Papua New Guinea

Nutrients in kg haÀ1

Site

Hobu



Plant part

Marketable tubers

Nonmarketable

tubers

Vines



Fresh yield

(Mg h1)



N



P



K



18.2 Ỉ 3.7

4.0 Ỉ 1.0



30 Ỉ 6

8Ỉ2



12 Æ 2

3Æ1



93 Æ 20

25 Æ 6



26.2 Æ 4.8



80 Æ 8



18 Æ 2



180 Æ 30



61 Æ 13 20 Æ 2



33 Æ 3 298 Ỉ 46



67 Ỉ 12 26 Ỉ 2



118 Ỉ 10



Total

Unitech Marketable tubers

Nonmarketable

tubers

Vines

Total



9.0 Ỉ 3.8

2.9 Ỉ 1.3

30.1 Ỉ 8.2



15 Ỉ 17

5Ỉ5



Ca



Mg



5Ỉ1

5Ỉ1

1 Ỉ 0.5 1 Ỉ 0.5



7Ỉ3

2Ỉ1



39 Ỉ 19

12 Ỉ 5



4Ỉ2

2Ỉ1

1 Æ 0.5 1 Æ 0.5



590 Æ 21 22 Æ 2



189 Æ 15



37 Æ 8



79 Æ 40 31 Æ 5



241 Æ 23



42 Ỉ 10 13 Ỉ 3



10 Ỉ 2



Hobu soils were classified as Typic Eutropepts and the soils at Unitech were Typic Tropofluvents

(Hartemink et al., 2000).



In annual crops only part of the total amount of nutrients taken up is

removed by the economic produce viz. the grain of wheat, the tubers of sweet

potato, or the seeds of soybean. An important portion of the nutrients taken

up may be returned to the soil with the cropping residues. Table IX gives the

total nutrient uptake of sweet potato tubers and vines (¼ above‐ground

biomass); less than one‐third is found in the marketable tubers (¼ economic

produce). Farmers only remove the tubers from the field and the vines

remain behind as crop residues. As vines decompose, nutrients become

available for the subsequent crop. Less than 25% of the total N and K

uptake was found in the economic produce. It is generally recognized that

crop residues are extremely important for recycling of nutrients in many

cropping systems in the tropics (Giller et al., 1997; Kumar and Goh, 2000).



VI. PRESENTATIONS OF RESULTS

For Type I data (monitoring soil properties over time) it is essential that

the methods of soil analysis have not changed, that is, comparing soil

organic C determined by the Walkley & Black method in 1970 to values

obtained from the same field using a dry‐combustion analyzer in the year

2000 is less than ideal. Provided analytical methods are unchanged, simple

t‐tests or analyses of variance can be used to detect statistically significant

diVerences. An example of Type I data are given in Table X, whereas Type II



204

Table X

Topsoil Chemical Properties of Fluvents and Vertisols Between 1979 and 1996 (Arithmetic Mean Ỉ 1 SD) of a Sugarcane Plantation in Papua New Guinea

Exchangeable cations (mmolc kgÀ1)



Vertisols



pH 1:2.5,

water



Organic

C (g kgÀ1)



Available

P (mg kgÀ1)



CEC pH 7

(mmolc kgÀ1)



Ca



Mg



1979

1982

1983

1984

1994

1996

1979

1982

1983

1986

1994

1996



15

14

44

9

12

8

6

17

40

7

12

12



6.5 Ỉ 0.4

6.2 Ỉ 0.1

6.3 Ỉ 0.1

6.1 Ỉ 0.1

5.9 Ỉ 0.1

5.8 Ỉ 0.2

6.6 Ỉ 0.1

6.2 Ỉ 0.1

6.3 Ỉ 0.2

6.2 Ỉ 0.2

5.9 Ỉ 0.1

5.8 Ỉ 0.2



58 Ỉ 15

nd

nd

nd

35 Ỉ 6

31 Ỉ 7

52 Ỉ 9

nd

nd

nd

32 Ỉ 3

32 Ỉ 6



nd

36 Ỉ 4

37 Ỉ 10

42 Ỉ 10

28 Ỉ 9

28 Ỉ 12

nd

43 Ỉ 5

40 Ỉ 13

37 Ỉ 18

32 Ỉ 11

28 Ỉ 11



389 Ỉ 43

459 Ỉ 55

435 Ỉ 48

437 Ỉ 52

384 Ỉ 65

374 Ỉ 33

421 Ỉ 21

490 Ỉ 29

477 Ỉ 94

490 Ỉ 108

452 Ỉ 79

421 Ỉ 102



228 Ỉ 78

275 Ỉ 35

256 Ỉ 35

266 Ỉ 45

232 Ỉ 47

220 Ỉ 30

293 Ỉ 69

286 Ỉ 22

290 Ỉ 83

307 Ỉ 77

273 Ỉ 50

276 Ỉ 73



93 Ỉ 41

113 Ỉ 24

100 Ỉ 16

102 Ỉ 21

101 Ỉ 22

99 Ỉ 13

123 Ỉ 39

131 Ỉ 16

114 Ỉ 33

112 Ỉ 37

129 Ỉ 34

115 Ỉ 38



nd, no data.

Type I data modified from Hartemink (1998c).



Base

saturation

(%)



K

13.0 Ỉ

12.9 Æ

12.4 Æ

12.9 Æ

10.8 Æ

8.0 Æ

15.5 Æ

16.1 Æ

12.9 Æ

12.3 Æ

13.4 Æ

9.0 Æ



5.0

2.0

2.8

3.8

2.3

2.0

2.7

2.9

2.3

5.6

3.9

3.0



79 Æ 17

87 Æ 2

85 Æ 3

87 Æ 4

90 Æ 5

88 Ỉ 8

93 Ỉ 17

89 Ỉ 2

87 Ỉ 9

88 Ỉ 3

92 Ỉ 5

92 Ỉ 8



A. E. HARTEMINK



Fluvents



Year



Number of

samples



Oxisols

Land‐use systema

pH 1:2.5, water

Organic C (g kgÀ1)

Available P (mg kgÀ1)

CEC (NH4OAc pH 7) (mmolc kgÀ1)

Exchangeable Ca (mmolc kgÀ1)

Exchangeable Mg (mmolc kgÀ1)

Exchangeable K (mmolc kgÀ1)

Base saturation (%)

Exchangeable Al (mmolc kgÀ1)

Al saturation (% ECECc)



Ultisols



Psamments



Inceptisols on limestone



Bush

vegetation



Permanent

cropping



Bush

vegetation



Permanent

cropping



Bush

vegetation



Permanent

cropping



Bush

vegetation



Permanent

cropping



6.2

21

3

125

68

26

5

80

0

0



5.2

17

3

88

13

5

1

21

9

32



6.1

15

3

157

38

23

5

45

0

0



4.6

11

<0.5

110

11

5

3

24

nd





6.3

7

3

98

27

14

3

47

0

0



5.3

7

2

60

12

4

2

28

0

0



7.5

19

9

310

161

70

3

76

0

0



7.4

34

4

310

140

36

1

58

0

0



Type II data modified from Hartemink (1997b).

a

Sampled sites were within 100 m distance.

b

Aluminium saturation of the ECEC is calculated as: (Al/Ca ỵ Mg ỵ K ỵ Na ỵ H ỵ Al) Ã 100.

nd, no data.



ASSESSING SOIL FERTILITY DECLINE IN THE TROPICS



Table XI

Soil Analytical Data of Under Bush Vegetation and Permanent Cropping in Northeast Tanzania



205



206



A. E. HARTEMINK



data are given in Table XI. Both these comparisons allow for conclusions on

the eVects of continuous cultivation on soil chemical properties (Hartemink,

2003). There are other methods to use soil chemical data to assess soil

fertility decline including calculations on the rates of change and using

paired sequential samples—these methods are discussed in a later section.



A. RATES



OF



CHANGE



For each soil chemical property (w) measured over a given time span (t),

the following can be calculated:

the absolute diVerence: w1Àw2,

the change per year: (w1Àw2)/(t1Àt2),

and the rates of change in soil chemical properties: [(w1Àw2)/w1]/(t1Àt2) Â 100,

which gives the change in percentage per year of the initial level t1.

Very few studies have been conducted in which rates of change in soil

chemical properties were calculated. Calculating the rate of change in percentage per year using two data points assumes a linear change in a soil

property. However, many soil chemical processes are nonlinear and the rate

of change therefore diVers at diVerent periods (Jenny, 1980). For example, an

average decline in organic C at a rate of À0.3 g kgÀ1 yearÀ1 observed between

1980 and 2000 may have been À0.8 g C kgÀ1 yearÀ1 in the 1980s, but less than

À0.2 g C kgÀ1 yearÀ1 in the 1990s. This is further discussed in Section VII.

A diVerent method is to assume that loss of a nutrient, w, is a first order

kinetic process that can be fitted to single exponential model. The first order

process is:

dw=dt ¼ kt;

whereby the rate factor k can be calculated from plotting lnw/w0 versus

t whereby k represents the slope of the line. Calculating the k‐factor gives

insight in the rates of change in a property. This was suggested by Nye and

Greenland (1960) and first order kinetics have been widely used in crop

residue and organic matter decomposition studies. First order kinetics

were used by Arnason et al. (1982) in a study of soil fertility decline in

Belize: Table XII lists the results and shows the k‐factor and the relative

change in soil chemical properties of Rendols under permanent cropping in

Belize. For the use of the single exponential model, several data points and

relatively short time‐steps are needed. This method cannot be used when soil

fertility studies have only single time steps (t1 and t2). First order kinetics fit

well for C and N but less well for exchangeable cations or pH. Overall this

method provides necessary input for scenario studies on how soil fertility

changes over time.



ASSESSING SOIL FERTILITY DECLINE IN THE TROPICS



207



Table XII

Decline of Soil Fertility in Relative Values (Percentage per Year) and Calculated k‐factor

Based on First Order Kinetics

Soil chemical property



Relative rate of decline (% per year)



k‐factor (per year)



1.2

10

4.8

16

3.9



0.013

0.11

0.05

0.19

0.035



pH

Available P

Total N

Exchangeable Ca

Exchangeable K

Modified from Arnason et al. (1982).



B. PAIRED SEQUENTIAL SAMPLES

In some studies, several paired samples are available but all of them with

diVerent single time steps. This is the case when various fields are being

sampled at diVerent times, for example, some fields may have been sampled

in 1987 and again in 2003 whereas other were sampled in 1992 and again in

2000. The data set from such a sampling scheme has several values of a soil

property with diVerent time steps. It is possible to calculate from such data

the rate of change whereby the diVerence in years between the initial

sample (t1) and the second sample at (t2), is plotted against the diVerence

in the measured soil property values. Based on a large number of sample

pairs, the decline in a soil chemical property can be calculated whereby t1 is

the initial value and t2 the value of the second sampling. Thus, it can be

calculated whether a soil property had increased or not changed (i.e., value

at t2 minus value at t1 ! 0) or whether there has been a decline (i.e., value at

t2 minus value at t1 < 0).

From a sugarcane plantation in Papua New Guinea, pH data were

available from 80 fields sampled at diVerent sampling times. The diVerence

in years between the initial sample at t1 and the second sample at t2, was

plotted against the diVerence in the measured pH values. It appeared that the

decline in pH was related to the initial pH value (Fig. 1). Although the data

are scattered, a larger decline occurred when the initial pH was high. This

relation does not take into account the time elapsed between the pH measurements. Based on the 80 sample pairs, the decline in pH with time was

calculated whereby t1 was the initial value and t2 the pH value of the second

sampling. In only a few samples the pHw increased or had not changed (i.e.,

pH at t2 minus pH at t1 ! 0) but in the majority of the sample pairs there was

a decline in pH (i.e., pH at t2 minus pH at t1 < 0). The largest decrease in pH



208



A. E. HARTEMINK



Figure 1 Changes in topsoil pH (0–0.15 m) in relation to the initial pH at t1 (A), and the

change in topsoil pH with time (B). Based on 80 sample pairs. Type I data. Modified from

Hartemink (1998a).



occurred after 10 years (t2 À t1 ! 10), and nearly 50% of the variation was

explained by the linear function:

DpH ẳ 0:444 t2 t1 ị:

This method proved useful to quantify rates of change in a soil property

using paired sequential sample data with diVerent time steps.



C. BULK DENSITY

Cropping brings about changes in soil physical and soil biological properties and these also influence soil chemical properties. For example, changes

in the soil moisture or temperature regime aVect soil microbial biomass,

which influences mineralization of organic matter and other processes.

Measured changes in soil chemical properties are a net eVect of these processes but such changes also depend on the bulk density of the soil, which

may alter under cropping. In the previous section, changes in soil chemical

properties were mostly expressed as concentrations, for example, mmolc kgÀ1

or g kgÀ1. Nutrient concentration can be expressed as nutrient content

(kg haÀ1), which can be used in nutrient balance studies and translated in

nutrient replacement by inorganic fertilizers or other amendments.

Suppose an Alfisol cropped with millet contained 1.5 g N kgÀ1 in the

topsoils (0–0.20 m) in 1990, and 1.2 g N kgÀ1 in 2000. The rate of change in

total N content is 0.03 g N kgÀ1 yearÀ1. If the topsoil has a constant bulk

density of 1.3 Mg mÀ3, the decrease of 0.03 g N kgÀ1 yearÀ1 is equivalent to



ASSESSING SOIL FERTILITY DECLINE IN THE TROPICS



209



a loss of 78 kg N haÀ1 yearÀ1. This figure is easy to deal with, particularly

when it is expressed as inorganic fertilizer: the loss of N from the topsoil is

equivalent to 170 kg urea or 390 kg sulphate of ammonia. A further step is to

translate this nutrient loss into economic terms (Alfsen et al., 1997; Drechsel

and Gyiele, 1999; FAO, 2001a).

As this example showed, expressing soil chemical properties in kg nutrient

haÀ1 requires soil bulk density values, which are rarely measured in soil

fertility studies. Moreover, many soil chemical properties are determined by

an extraction method and the values are expressed in terms of availability.

Available means that the nutrient is susceptible to absorption by plants,

whereas availability means eVective quantity (Black, 1993). The amount of

available nutrients extracted may hold little relation with the total amount

of the nutrient in the soil and its availability over a given time span. The

availability aspect is irrelevant for C and N because total pools are measured.

Bulk density measurements thus improve the quantification of C and N loss

as it would be possible to relate N loss to the total N pools. For P, K, or Ca

that is not possible unless the total element concentrations were determined.



D. BULK DENSITY EFFECTS ON NUTRIENT STOCKS

Bulk density is likely to change under cropping, which has several eVects.

In annual cropping systems where no mechanization is used, increases in

bulk density are not so likely to occur. Increases may be caused by people

in the field or occur naturally; generally these increases are not spectacular.

In mechanized annual cropping systems, where tractor traYc is common,

substantial compaction may occur (Soane, 1990), which may aVect the outcome of nutrient stock calculations. It may severely reduce nutrient availability (Arvidsson, 1999; Lipiec and Stepniewski, 1995) because rooting is

restricted which limits the volume of soil from which nutrients can be extracted. In compacted soils the crop also becomes susceptible to water stress,

which may have a larger impact than the reduced nutrient availability. It is

diYcult to distinguish these factors and their eVects on crop productivity.

When the soil is more compacted the thickness of the layer decreases. This

means that, if the sampling depth remains the same, part of the subsoil is

being sampled, which aVects calculations on nutrient contents. So sampling

should be corrected for decrease in the thickness of the compacted layer

(Dias and NortcliV, 1985).

An increase in bulk density does not mean that nutrient content is

reduced. Table XIII shows the nutrient concentration and nutrient content

of an Oxisol cropped with sugarcane. The nutrient content was calculated

for three depths using bulk densities determined in 1978 and 1983. Absolute

and relative diVerences in the nutrient concentration and nutrient content



210



Table XIII

Nutrient Concentration and Nutrient Content of Oxisols Under Sugarcane in 1978 and in 1983

Nutrient

concentration

Sampling

depth (m)

0–0.12



0.70–0.80



Soil property



1978



1983



Absolute



Percentage



BD (Mg mÀ3)

Organic C (g kgÀ1)

Total N (g kgÀ1)

Total P (g kgÀ1)

Ca (mmolc kgÀ1)

Mg (mmolc kgÀ1)

K (mmolc kgÀ1)

BD (Mg mÀ3)

Organic C (g kgÀ1)

Total N (g kgÀ1)

Total P (g kgÀ1)

Ca (mmolc kgÀ1)

Mg (mmolc kgÀ1)

K (mmolc kgÀ1)

BD (Mg mÀ3)

Organic C (g kgÀ1)

Total N (g kgÀ1)

Total P (g kg1)

Ca (mmolc kg1)

Mg (mmolc kg1)

K (mmolc kg1)



0.76

68.2

4.0

1.1

29.0

2.9

3.0

0.86

7.1

1.2

0.9

1.6

0.5

0.6

1.01

3.6

0.5

1.1

2.0

0.4

0.3



1.02

41.3

1.9

0.9

9.1

1.6

1.2

1.06

10.9

1.0

1.1

4.0

0.4

0.5

1.10

3.0

0.5

1.1

2.2

0.2

0.5



ỵ0.26

26.9

2.1

0.02

19.9

1.3

1.8

ỵ0.20

ỵ3.8

0.2

ỵ0.02

ỵ2.4

0.1

0.1

ỵ0.09

0.6

0

0

ỵ0.2

0.2

ỵ0.2



ỵ34

39

53

18

69

45

60

ỵ23

ỵ54

17

ỵ22

ỵ150

20

17

ỵ9

17

0

0

ỵ10

50

ỵ67



Calculated from data in Masilaca et al. (1985).



Nutrient content

(kg ha1)



DiVerence



1978



1983



Absolute



Percentage



62,198

3,648

1,003

530

32

107



50,551

2,326

1,102

223

24

57



11647

1322

ỵ99

307

8

50



19

36

ỵ10

58

26

46



6,106

1,032

774

28

5

20



11,554

1,060

1,166

85

5

21



ỵ5448

ỵ28

ỵ392

ỵ57

0

ỵ1



ỵ89

ỵ3

ỵ51

ỵ208

0

ỵ3



3,636

505

1,111

40

5

12



3,300

550

1,210

48

3

22



336

ỵ45

ỵ99

ỵ8

2

ỵ10



9

ỵ9

ỵ9

ỵ20

46

ỵ82



A. E. HARTEMINK



0.300.40



DiVerence



Tài liệu bạn tìm kiếm đã sẵn sàng tải về

B. Nutrients in the Roots and Crop Residues

Tải bản đầy đủ ngay(0 tr)

×