Table 1. Average rates of nitrogen fertilizer used
on arable and grassland in England and Wales.
Source: Mattikalli (1993).
Year N (kg/ha)
Arable Grassland
1. 1943 19 4
2. 1962 74 23
3. 1970 90 53
4. 1984 200 60
5. 1987 200 60
6. 1989 200 60
2. THE EXPORT COEFFICIENT MODEL
The export coefficient model estimates
nitrogen loading in the surface water at the outlet
of a watershed using areal extents of different
types of land-use within the watershed and the
associated fertilizer application rates (Reckhow
and Simpson, 1980; Beaulac and Reckhow, 1982;
Johnes and O'Sullivan, 1989). The model follows
from that developed by Jorgensen (1980) which
calculated the areal external nutrient load upon a
water body by summing outputs from all
potential sources. Using this concept, the nitrogen
losses from a watershed to surface water are
estimated for all possible sources of nitrogen, and
by defining appropriate export coefficients. As
the present study is focused on a predominantly
rural watershed, only losses from arable,
grassland, and woodland are considered.
Sediment losses are not addressed owing to the
lack of data in the literature on sediment export
rates.
The nitrogen loss, L (in kg/year), at the
outlet of a watershed is given by:
n m
Zt Yn]: a
i=1 £5)
= area of a discrete land-use patch (ha)
E = export coefficient for a specific land-use
type (kg/ha)
m = number of discrete patches under one land-
use class
n = number of land-use classes
a = artificial inputs (e.g. sewage, industrial etc.)
(kg/year)
134
An export coefficient (E) indicates the .
proportion of the total nitrogen added to a given
land-use patch that is lost to surface water. This is
the key parameter, which determines the amount
of nitrogen loading from the various land-use
types, and is obtained from the literature.
Individual coefficient (such as Ewoodland:
Egrassland, Earable €tc) is deemed to allow for
biological, atmospheric and geological
(weathering) contributions to nitrogen loss from
each land-use type. This is a modification
compared to the original structure of the model;
similar modifications were employed in the
version constructed for Slapton Ley by Johnes
and O'Sullivan (1989). The parameter values and
export coefficients used in this application of the
model were compiled from the literature (Table
2). Nitrogen input and loss from the human and
livestock population are considered here because
urban area constitutes a very small proportion in
the study area and the emphasis of the this study
is only on the agricultural inputs.
Table 2. Export coefficients for nitrogen. Source:
Mattikalli (1993).
Land-use N fixed N loss
(%) (kg/ha)
Temporary grass 5 30
Permanent grass S 50
Cereals 12 4
Root crops 20 10
Oilseed rape 30 4
Woodland 13 20
Rough grazing and Set-aside 2 10
3. STUDY AREA
AND DATA DESCRIPTION
The River Glen watershed in eastern
England is selected. The geographical co-
ordinates of the area are: longitude 09 20' - 09
40' W; latitude 529 40' - 539 00' N, and it extends
between Grantham, Bourne and Stamford in
South Lincolnshire. This region is characterized
by major changes in agricultural activity with
increased arable farming in the past few decades.
The choice of study area was influenced by the
existence of considerable quantities of land-use
data required by the export coefficient model.
The study area is mainly used for agricultural
purposes, and other land-uses such as urban areas
form a small proportion. Crops such as wheat,
barley and sugar beet dominate the watershed. In
addition to arable farming, grassland cultivation is
also important within the watershed, although
much of the grassland in the region has changed
to arable.
data sets
quality.
are neces
validate n
His!
are availa
data are a
1st Land
scale of
survey, al
of air p
Remotely
derive lar
The 1984
MSS ima
while the
HRV ima,
1989, lan
Landsat '
obtained 1
(ITE), Mc
for 1989
map of t
Environm
1994). Th
enormous
scale and
Dai
available
(NRA) fo
clear fror
Coggles,
stations h
with land
nitrogen e
of the rec
Nitrogen
Oxidized :
to 1981 i
nitrites, ar
forms of 1
to comp
consistenc
AG
for the m
volumes c
vector-bas
since the 1
copy pap
digital sc.
input land
in geomet
than 7 m