Soil Test and Agricultural
Management Practices Survey
of the Clear Lake Watershed
Antonio Mallarino and Jeremy Klatt
A. Introduction
This
work was a component of the project Clear Lake Restoration Diagnostic and
Feasibility Study lead by Dr. John Downing (Department of Animal Ecology),
which included other components and co-investigators. Objectives and methods of the overall project, and how the
component fits with the overall project objectives were described in the
original project. The objective of this
component was to survey surface soil test values, with emphasis on phosphorus
(P), and nutrient management practices within the agricultural area of the
watershed. This information is a key
aspect that should help understand and assess the role of nutrient levels and
management practices on the nutrient load to the lake and its water quality
status. This knowledge will help
determining alternatives for improving the water quality of the lake.
B. Methods
The
work included a survey of nutrient levels in soils that involved extensive soil
sampling and a survey of relevant soil and crop management practices. Plans for both surveys were developed during
summer and early fall of 1999 in cooperation with personnel of the CLEAR Project,
in particular with Mr. Ric Zarwell, who was CLEAR Project Coordinator until
December 1999. Efforts until early
September were directed toward developing the plan for the survey and to
request permission from landowners to collect soil samples from their
fields. This part proved to be more
difficult than expected, and finally permission was denied for approximately
20% of the fields. Except for the small
agricultural area at the extreme southeast corner of the watershed, however,
the information collected provides a good representation of nutrient status and
agricultural practices in the watershed.
1. Management Practices Survey. A
two-page questionnaire was mailed or delivered in person to all landowners or
operators from whose fields soil samples were collected in order to obtain
information of agricultural practices that are occurring in the watershed. The questionnaire requested information
about crop rotation, tillage system, and rates of application and type of P
fertilizer, nitrogen (N) fertilizer and manure used during the last five
years. Sixty-six percent of the
surveyed farmers responded, which represents approximately 60% of the land in
the watershed. This success rate is
typical of these types of surveys (when no payment is included), and this high
rate was achieved only after repeated telephone calls and some personal visits.
2. Soil Nutrients Survey. The
soil sampling strategy was planned based on available information on field
borders, soil types, current crop and planned fall fertilizer management. Consideration of the objectives,
characteristics of the watershed, previous information of areas from where
seemingly much of the P load to the lake could be derived from and the
available budget were key aspects for deciding among several soil sampling
alternatives. The sampling plan chosen
was a systematic stratified sampling method, in which predominant soil map
units of each field were sampled following a point mapping and sampling strategy. In a first step, approximately 500 small
sampling areas (points) about 0.2-acre in size were identified within the
predominant soil map units within each field on the basis of digitized and
georeferenced soil survey maps, georeferenced maps of field borders, surface
drainage, aerial photographs and visual observations of the watershed. Geographical Information Systems (GIS)
techniques and ArcView computer software were used to produce several layers of
information. A second step involved a
field verification and modification (when necessary) of the location of the
sampling points through use of hand-held Global Positioning Systems (GPS)
receivers and field observations. The
original number of 500 desirable sampling points had to be reduced to approximately
340 due to several reasons. These included
lack of permission to sample several fields, that a small number of fields
could not be sampled because contrary to our request the farmers applied P
fertilizer immediately after harvest and before samples could be collected, and
because of budget limitations. In a
third step, one composite soil sample made up of 15 to 20 soil cores were
collected from the top 6 inches of soil of each sampling area. This is the sampling depth recommended by
ISU for agronomic soil tests. The GIS
map in Figure 1 shows the sample points location.
All
soil samples were dried, ground to pass a 2-mm screen, and analyzed in
duplicates by several agronomic soil tests supported by Iowa State
University. These included analyses for
Bray-1 P (which will be referred to as the Bray test hereon), Olsen (or sodium
bicarbonate) P, Mehlich-3 P (which will be referred to as M3 test hereon),
ammonium acetate-extractable K, Ca, and Mg, soil pH and organic matter. The Bray test is the most commonly used P
test by soil test laboratories and farmers in Iowa, and the Olsen P and M3 P
tests provide a different assessment of soil P, mainly in high-pH soils. In addition, all soil samples were analyzed
by the two environmental P tests most commonly used in ongoing research across
the United States. These type of tests
are the object of intensive research in Iowa and other states because they
could provide a different assessment of the potential impact of soil P levels
on P losses with water runoff or through the soil profile. Agronomic soil P tests were developed and
calibrated to predict P sufficiency for crops, not necessarily to predict P
losses to water resources. The tests
chosen were the iron oxide-impregnated paper P test (which will be referred to
as the Iron-oxide test hereon) and the deionized water extraction test (which
will be referred to as the Water test hereon).
The procedures used are not described here, but information can be found
in published articles. Procedures
followed for all agronomic tests were those described in the North Central
Region Publication 221 (Brown, 1999), which includes agronomic soil test
procedures recommended by the North Central Region Committee for Soil Testing
and Plant Analysis (NCR-13). Procedures
followed for the two environmental P tests were those described in the Southern
Cooperative Series Bulletin 396 (Pierzynski, 2000), which includes procedures
for a variety of P tests recommended by the national committee Minimizing
Phosphorus Losses from Agriculture - Information Exchange Group (SERA-17/IEG).
When
soil-test P values were summarized into GIS maps, values of agronomic soil P
tests were grouped into the agronomic soil test interpretation classes used by
Iowa State University. Two
modifications were made, however, to better describe the results. One was that the Very Low and Low classes
were merged because very few fields tested Very Low and little distinction can
be made between these two classes for the purpose of environmental
conservation. The other modification
was to split the Very High class into two classes to better describe the
high-testing soils. The Optimum class
(16 to 20 ppm for the Bray or M3 tests and 10 to 14 ppm for the Olsen test)
defines soil test values for which only maintenance fertilization is
recommended to account for P removal in harvested products. No P fertilizer is recommended for the High
or very High classes by Iowa State University.
3. Data Management Procedures. The
procedure used to summarize the soil P and field history survey information was
to summarize the information for various sub-watersheds delineated according to
surface drainage patterns. Regression
and correlation techniques were used to study how the various soil test
extractants assessed soil P.
C. Summary Results
1. Field Management Practices. Summaries
of relevant field history are shown in Table 1 and several figures. The Clear Lake watershed is dominated by
row-crop agriculture. Eighty-nine
percent of the acreage surveyed was in a corn and soybean rotation (Fig.
2). Of the remaining 11% of the land,
7% was being cropped for corn continuously and the remaining land was either
set aside as part of the Conservation Reserve Program (CRP) or under alfalfa or
pasture.
The
most common tillage practice in the watershed is a tillage system that includes
chisel plowing in combination with disking and/or field cultivation. Forty-eight percent of the land was managed
with this tillage system (Fig. 2). The remaining
fields under row-crop production were split between no-till (21%), ridge-till
(18%), moldboard plow (7%), and those that include V ripping (6%). From a soil conservation perspective, the
moldboard plow system is the most prone to produce soil erosion and the no-till
system usually produces the lowest soil erosion. The other tillage systems usually rank intermediate and their
impact on soil erosion is highly dependant on the residue cover.
The
most common P application method in the watershed is to incorporate the
fertilizer into the soil in the fall by plowing, disking, or injecting. Forty-four percent of the land had P applied
in this way. The next two most common
methods were broadcast application in the fall without incorporation (13%) and injection
or incorporation of the fertilizer into the soil in the spring (4%). Fall application with incorporation is the
preferred method for P application from an environmental perspective as most
runoff events occur in the spring.
Injecting or incorporating is preferred over leaving it on the soil
surface because this reduces the risk of P loss with surface runoff. When a no-till system is used, however, the
advantage of incorporating P fertilizer or manure by injecting into the soil
over a surface application with good conditions (other than application to
frozen, water saturated, or snow-covered ground) is not clear, however. Commonly used injection equipment cause
significant soil disturbance, which may increase soil erosion and total P loads
to surface water resources. Soil loss
(and loss of P attached to soil particles) is very small for no-till or
pastureland that receives broadcast fertilizer or manure applications but loss
of dissolved P can be high when soil conditions are adverse.
On
average, during the last five years P fertilizer was applied 2.2 times. When P fertilizer is applied, it is
typically applied at a rate of about 65 lb P2O5/acre. For comparison, Iowa State University
recommends 70 lb P2O5/acre be applied once every two
years to soils testing in the optimum range in a corn and soybean
rotation. The Optimum range for the
Bray or M3 soil P tests is 16-20 ppm.
Consideration of the average M3 soil test P value of 40 ppm and the
average application rate of 65 lb P2O5/acre may suggest
that the watershed is being over fertilized, which is typical of most agricultural
regions of Iowa. However, a close look
at specific areas of the watershed shows that some fields testing much higher
than the average are also being fertilized.
These observations suggest that many farmers in the watershed are
managing their land using recommended rates of application and that only a few
are over-fertilizing their fields. This
aspect will be visited further when results of the soil sampling are discussed.
Only
4% of the land received fall-applied N.
In the last several years there has been a push to reduce fall
application of N on agricultural lands.
Nitrogen, unlike P, is very mobile in the soil and under the right
conditions can be lost with surface runoff or leach through the soil profile
into sub-surface drains or to groundwater supplies. The most common application method was broadcasting without
incorporation in the spring (50%), followed by spring sidedressing (26%), and
injection (11%). The remaining land had
no N applied (8%). On average in the
last five years, farmers have applied N 2.4 times to their fields. Farmers, for the most part only apply N
before corn due to the ability of soybeans and forage legumes to fix
atmospheric N. When applied, it is
typically applied as anhydrous ammonia at a rate of about 130 lb N/acre.
Our
survey showed that 18% of the farmland had received manure in the previous five
years. The responses in most cases
provided insufficient data concerning rates of manure application. Manure often can be overapplied to certain
areas in order to dispose of it, which creates high nutrient levels in
concentrated spots. The data actually
shows, however, very similar average soil P tests for manured fields and
nonmanured fields in this watershed.
The average M3 soil test P value of manured soils was 34 ppm while
nonmanured soils averaged 40 ppm.
Manure is also thought to create more variability within a field but our
data suggests very similar variation of P in manured and nonmanured soils. Our numbers showed higher organic matter
percent and calcium in fields where manure was frequently applied.
2. Soils of the Watershed. The
Clear Lake watershed is dominated by loam and silty-clay loam textured soils
under a predominately agricultural setting.
The soil survey map is shown in Figure 5. Background chemical analyses of the collected samples that are
useful to describe the fields include organic matter, and extractable K, Ca,
and Mg. Mineral soils predominate and
have an average organic matter content of 5.7%. According to the digital soil survey maps, these soils typically
have a cation-exchange capacity ranging from 0-65
(meq/100 gm) with an average
of about 22 (meq/100 gm). The average
pH in the watershed was 6.6. Calcium
and Mg concentrations averaged 3357 ppm and 380 ppm respectively. The K values, which ranged from 49 ppm to
502 ppm and averaged 170 ppm, indicate that there is extensive but very uneven
K fertilization occurring in the watershed.
The optimum range for K for corn and soybeans as recommended by Iowa
State University is 90-130 ppm. The
landscape forms, soil types, and drainage patterns, which have strong influence
on nutrient delivery to the lake were considered together with soil-test P data
in others components of the project.
3. Soil Test Phosphorus.
a. Soil P assessment by five soil P tests. There were
differences in amounts of P extracted by the five soil P tests used. Differences or similarities between tests
and reasons for any difference should be discussed before interpreting
distribution of soil-test P across the watershed. Average amounts of P extracted by the five methods are shown in
Table 2. The M3 soil test extracted the
highest amount of P followed very closely by the Bray test in most samples
(except in those with high pH). Iowa
research has shown that this test usually extracts similar to or only slightly
higher amounts of P than the Bray test, except when soils have high pH and
calcium carbonate content. This result
was confirmed in this project because the Bray extracted much less soil P than
the M3 in soils with pH 7.2 or higher.
The Olsen and Iron-oxide test extracted very similar amounts of P that
were less than the amount extracted by Bray or M3 tests. The Water P test extracted the lowest amount
of P. All soil tests used with
agronomic or environmental purposes extract only a small proportion of the
total soil P, which includes the P fraction assumed to be better related to
soil P sufficiency for crops (for the agronomic tests) or to potential loss of
P dissolved in water (for the environmental tests). Current Iowa research is developing field calibrations for these
tests and others in relation to potential loss of P to water resources. The ongoing research has not determined
clear differences between the tests, and this is the main reason these five
tests were chosen for this survey.
Three
of the five soil P tests used were highly correlated across all samples
collected (Table 3). The correlation
coefficients between the M3, Olsen, and Iron-oxide P tests were very high and
ranged from 0.96 to 0.97 across samples.
Correlations involving the Bray and Water P tests were variable and
usually lower. Correlations
coefficients for the Bray test, the most commonly used agronomic P test in
Iowa, were the lowest and ranged from 0.88 with the Water test to 0.92 with the
M3 test. Correlations coefficients for
the Water test (an environmental P test) were intermediate and variable, and
ranged from 0.91 with the Bray test to 0.96 with the Iron-oxide test (which is
another environmental P test). These
differences in correlations between tests may suggest differences concerning
estimates of P loss from the fields with surface runoff and tile drainage. While the Bray, M3 and Olsen tests are used
for agronomic purposes, the environmental tests would predict better potential
P delivery to surface water resources.
The correlation between the Bray soil test and the other tests across
all samples was poorer because this test extracted proportionally less P from
soils with pH higher than 7.2, which likely were calcareous. This is a common occurrence observed in Iowa
soils. Of the 332 samples analyzed from
the watershed, 109 had pH higher than 7.2 and tended to have the highest exchangeable
calcium concentrations. Correlation
coefficients in Table 3 that included only soils with pH less than 7.2 showed
much higher correlations for the Bray test.
The intermediate and variable correlations for the Water test either
across all soils or for noncalcareous soils suggest that this test extracts
slightly different P fractions than other tests from different soils.
Overall,
the only test that clearly departs from the others was the Bray test in high pH
soils. This result may indicate that
the Bray test should not be used across all fields in this watershed for agronomic
assessment because it would underestimate sufficiency levels for crops. Conclusions concerning assessments of the
potential P loss to the lake are not so clear cut, however. The Bray test would underestimate the amount
of P tied-up to soil particles that could reach the lake in eroded soil. However, the presence of calcium carbonate
could reduce P that could be released from the soil phase of the soils and that
could be transported dissolved in water though surface runoff or subsurface
drainage. If this were the case, the
Bray test would provide a more accurate assessment of dissolved P loss. Theoretical considerations suggest that the
two environmental tests would be less affected by soil properties and would
provide better assessments of the potential P loss across all conditions. The soil test correlations suggest, however,
that the Iron-oxide, M3, and Olsen tests provided similar environmental
assessments which differed slightly with assessments obtained by using the Bray
and Water tests.
b. Distribution of soil-test P throughout the
watershed. There
was a wide range of soil test P values throughout the watershed. Data for the individual soil samples
collected from each field showed that values for the M3 P test (for example)
ranged from 2 to 395 ppm across the watershed while values for the water P test
ranged from 1 to 68 ppm (not shown).
These wide ranges of values confirm that there has been extensive P
fertilization occurring in the watershed.
Available soil P of most Iowa soils was in the Very Low or Low
interpretation classes until approximately the mid 1900s, and were increased to
current levels by fertilization and/or manure management practices. The average M3 P test across the watershed was
40 ppm while the median value was 32 ppm.
The median may be a better value to look at to prevent a few very high
soil samples from skewing the data. The
median value is higher but not very much higher than the value considered
optimum for agronomic purposes (20 ppm).
Also, although this median value is high, it is not far from the range
that could be considered acceptable (16 to 25 ppm) if the usually very high
variability in soil tests due to high and largely unavoidable sampling error is
considered. Iowa research shows that a
soil test value of about 30 ppm would drop into the Optimum level within two to
three years of cropping without P fertilization.
The
distribution of soil test values for sub-watersheds is shown in several GIS
maps. On a field basis, using the M3 P
test as an example, soil P ranged from 6 to 296 ppm across the fields of the
watershed and averaged 41 ppm. Several
fields tested very high in P.
Approximately one third of the fields tested Very High by any of the
three agronomic soil tests. The
high-testing fields were spread without clear patterns across the
watershed. In spite of lower
correlations discussed above for the Bray and Water P tests, the five tests
used tended to agree in identifying high-testing fields.
The
other approach used for summarizing the soil test information was to calculate
averages for sub-sections of the watershed delineated according to surface
drainage patterns. The results of using
this approach are shown in the GIS maps in Figures 4 to 9. This approach provides a slightly different
perspective of the distribution of soil P and better idea of the potential
impact of current management practices in different parts of the watershed on P
loads to the lake and, also is useful to identify sub-watersheds where changes
in P or soil conservation practices are most urgently needed. Continued use of the M3 P test as an example
shows that soil P levels ranged from 18 to 65 ppm across all sub-sections of
the watershed. In spite of minor
differences between the five P test used, all tests identified high-testing
subsections in the western side of the watershed, and also three smaller
high-testing subsections bordering the south, southeast, and northeast sides of
the lake. Although these maps must be
interpreted with caution because each sub-section was highly heterogeneous, the
results suggest the areas that potentially contribute a large proportion of the
P loads to the lake.
The
survey of soil P status and P management practices of the Clear Lake
agricultural watershed was useful to identify areas that may be sources of
large P loads to the lake and to identify priority areas where changes in P
management practices would be desirable.
Approximately one third of the area of the watershed had soil-test P
values that were twice to five times higher than levels needed to maximize crop
production. The highest soil-test P
values were found in a very small number of fields that received either P
fertilizer or manure, which likely are the source of a major proportion of the
P being transported by surface runoff or subsurface drainage to the lake. There were minor differences between soil
tests in describing the distribution of soil P across the watershed. All tests identified high-testing
areas. The Bray test measured proportionally
less P from high pH soils, and its use may mislead producers to apply more P
fertilizer than needed to optimize crop production.
Phosphorus
fertilizers or manure are not incorporated into the soil in approximately 30%
of the area, a major proportion of which is under no-till or ridge-till
management. Only approximately 40% of
the area surveyed is under no-till or ridge-till tillage systems, which are the
most effective in reducing soil erosion and surface water runoff. The finding that several fields had
above-optimum soil P levels for crop production and were managed with
conventional tillage do not necessarily mean that these fields are major
sources of P loads to the lake.
However, the results strongly suggest that further adoption of commonly
recommended best management fertilization and soil conservation practices by
some producers would have a major impact in reducing the risk of P delivery
from the agricultural area of the watershed.
References
Brown, J. R. (ed.).
1998. Recommended soil test
procedures for the north central region, North Central Regional Research
Publication No. 221 (Revised).
Pierzynski, G.M. (ed.). 2000. Methods of
phosphorus analysis for soils, sediments, residuals, and waters. Southern Cooperative Series Bulletin No. 396
Table 1. Summary of the field history information.
|
Acres |
% of land |
Primary tillage |
|
|
Chisel/Disk |
1983 |
48 |
Moldboard plow |
291 |
7 |
No-Till |
844 |
21 |
Ridge-Till |
725 |
18 |
Other or unknown |
229 |
6 |
|
|
|
Cropping System |
|
|
Alfalfa, pasture, or CRP |
195 |
4 |
Continuous corn |
266 |
7 |
Corn-soybean rotation |
3649 |
89 |
|
|
|
Method of P
Fertilization |
|
|
Fall injection/incorporation |
1766 |
43 |
Spring injection/incorporation |
687 |
17 |
Fall left on top |
706 |
17 |
Spring left on top |
532 |
13 |
No P applied |
349 |
8 |
Unknown |
70 |
2 |
|
|
|
Method of N
fertilization |
|
|
Fall injected/incorporated |
177 |
4 |
Spring injected/incorporated |
447 |
11 |
Spring sidedress |
1068 |
27 |
Spring left on top |
2069 |
50 |
No N applied |
349 |
8 |
|
|
|
Manure History |
|
|
No |
3330 |
81 |
Yes |
780 |
19 |
Table 2. Amount of P extracted by
five soil P tests and soil pH for 332 soil samples collected
throughout the Clear Lake Watershed.
Soil P test or pH |
Mean |
Median |
Range |
Minimum |
Maximum |
|
--------------------------- Soil test P (ppm)
-------------------------- |
||||
Olsen |
17 |
13 |
147 |
1 |
148 |
Bray-1 |
32 |
25 |
371 |
1 |
372 |
Mehlich-3 |
40 |
32 |
393 |
2 |
395 |
Iron oxide |
24 |
20 |
168 |
4 |
172 |
Water |
8 |
6 |
67 |
1 |
68 |
|
|
|
|
|
|
Soil pH |
6.6 |
6.5 |
3.4 |
4.8 |
8.2 |
Table 3. Correlation coefficients
between amounts of P extracted by five soil P tests from 332
soil samples collected throughout the Clear Lake Watershed.
Soil P test |
Olsen |
Bray-1 |
Mehlich-3 |
Iron oxide |
|
------------------- Correlation Coefficients
----------------- |
|||
Entire pH range of 4.8 to 8.2 |
||||
Bray-1 |
0.89 |
|
|
|
Mehlich-3 |
0.96 |
0.92 |
|
|
Iron oxide |
0.96 |
0.88 |
0.97 |
|
Water |
0.93 |
0.91 |
0.95 |
0.96 |
|
|
|
|
|
Samples with pH less than 7.2 |
||||
Bray-1 |
0.97 |
|
|
|
Mehlich-3 |
0.97 |
0.99 |
|
|
Iron oxide |
0.96 |
0.96 |
0.97 |
|
Water |
0.93 |
0.95 |
0.95 |
0.97 |