Illinois Fertilizer Conference Proceedings January 21-23, 2002 Home 2002 Index Search

Real-Time, Web-Based Dissemination of Illinois Soil Temperature

Steven E. Hollinger and Robert W. Scott¹

Introduction

Fall application of nitrogen (N) fertilizer is a common practice in Illinois because wet soils in the spring make timely application of N and other spring fieldwork highly uncertain. Any delays due to wet soils may result in delayed corn planting, with accompanying yield decreases. Soil moisture is usually recharged in Illinois during fall and early winter. When the soil moisture is fully recharged, field operations are generally halted until spring fieldwork begins in late March or early April. Early fall application of N, when the soils are warm, results in large losses of the nutrient and contributes to the degradation of water quality. To limit these losses, fall application of N should be delayed until after the 4-inch soil temperature is less than 60°F (15°C) if a nitrification inhibitor is applied, and after the soil temperature is below 50°F (10°C) if no inhibitor is applied (Hoeft, 1998).

The median date when the 4-inch soil temperature falls and remains below 60°F during autumn varies from 12 October in the northern part of Illinois to 1 November in the south (Figure 1). For a soil temperature of 50°F, the dates range from 1 November in the north to 26 November in the south. These dates occur five years out of 10 and are the last occurrence of a 60°F or 50°F 4-inch soil temperature in the autumn. In one year out of 10, the last 60°F temperature can occur as early as 22 September in the north and 12 October in the south. For a 4-inch soil temperature of 50°F, the corresponding dates are 17 October in the north and 6 November in the south. In nine years out of 10, the last 60°F occurs as early as 27 October and as late as 26 December; the last 50°F temperature occurs as early as 11 November and as late as 31 December.

Because of the wide range in the dates on which the last 60°F or 50°F 4-inch soil temperature occurs, it is necessary to use actual soil temperature measurements to determine when to begin fall application of N. The preferred soil temperature measurement is at a 4-inch depth under bare soil at 10 a.m. CST. Tracking the soil temperature changes requires daily measurements.

The primary objective of this project was to provide near real-time bare soil temperature data for Illinois to support timing of fall application of N fertilizer.

Materials and Methods

The Illinois State Water Survey (ISWS) maintains a network of automated weather stations throughout the state (Hollinger et al. 1994). The 18 stations that make up the network are located at research and education centers or community colleges (Table 1). These stations make up the Illinois Climate Network (ICN). The variables measured include hourly air temperature; relative humidity; precipitation; solar radiation; wind speed and direction; barometric pressure; soil moisture at 2, 8, 20, 40, and 60 in.; and soil temperature at the soil surface and at approximately 1, 2, 4, 8, 12, 16, 20, 24, 28, 32, 36, and 40 inches. The soil variables are measured under a grass cover; therefore, the values reported may be inaccurate estimates of bare soil temperature.

Baker (1965) reported that soil temperatures under sod had a smaller diurnal temperature range than bare soil temperatures. In July, the bare soil temperature was warmer than soil temperature under sod. In January, soil temperatures under sod were warmer than bare soil temperatures. Rodskjer and Tuvesson (1976) reported autumn mean temperature under newly sown wheat (bare soil) was about 1°C cooler than under short grass. From these measurements, it is expected, during the summer and winter, that 4-inch soil temperatures under sod should be different from bare soil temperatures. During the autumn (spring) when the soil is cooling (warming), temperature differences should also occur. However, because the sod temperatures are warmer in the winter and cooler in the summer than bare soil temperatures, there should be a time in both the spring and autumn when the bare soil and sod soil temperatures should be equal.

An unknown was whether the bare soil temperatures would be warmer or cooler in the autumn than the soil temperatures measured under sod. Therefore, soil temperature sensors were installed under bare soil plots at a depth of 4 inches in close proximity to temperature sensors at 4 inches below a sod surface. The bare soil temperature sensors were installed at Brownstown, Bondville, De Kalb, Dixon Springs, Orr, Rend Lake, Sand Farm near Kilbourne, and Stelle. These stations form north to south and east to west transects across the state. The stations were selected so that all major soil types were represented (Table 2).

Soil temperatures at a depth of 4 inches under sod and bare soil conditions were measured continuously at these stations. These measurements were used to develop an adjustment for soil temperatures under sod so that they represent bare soil temperatures. Linear regression was used to determine the relationship between the maximum, minimum, and 10:00 a.m. CST daily bare soil and sod temperatures. The resulting equations were then used to estimate the 4-inch bare soil temperature at the remaining 11 ICN stations where bare soil temperature sensors were not installed. Independent relationships between the maximum and minimum 4-inch bare and sod soil temperature were developed using soil temperature data gathered at West Lafayette, IN from 1994 to 2000.

Weather data from the ICN stations is automatically measured at each station and retrieved each day. To ensure high quality data, each day's data are checked for problems due to sensor or equipment failure. Computer programs were written to automate the quality control and archival of all weather data being collected by the ICN stations. Quality control procedures consist of evaluation of collected data to identify erroneous information and prevent it from being added to the archived data (Hollinger et al., 1994).

A structured website was developed to disseminate the soil temperature and weather data from the ICN to the public. The website includes the ability for users to see maps, graphs, and tables of current and historic soil temperature and weather data.

Results and Discussion

The initial soil temperature website, http://www.sws.uiuc.edu/warm/soiltemp.asp, became available in mid-October. This website is updated daily and provides maps of the 10:00 a.m. CST 4-inch bare soil temperature on the previous day, along with the previous day's maximum and minimum 4-inch bare soil temperatures. A six-day history of soil temperatures is also available. Included on the opening page of the soil temperature website is the following description and caution:

In addition to the maps of the sod and bare soil temperature, maps of the maximum and minimum air temperature, mean relative humidity, total solar radiation in MJ m^-2 d^-1, precipitation, mean wind speed, and potential evapotranspiration are available for the previous day. Hourly and historical data for a station can be viewed in two-dimensional graphical form.

Sod and bare soil temperature relationship

Two independent data sets were used to develop a relationship between the 4-inch bare soil temperature and the 4-inch sod temperature. The first method involved data collected as a part of this project, and the second involved historical data from Champaign-Urbana using the 4-inch bare soil and sod temperatures collected at the West Lafayette, IN weather station that is a part of the National Weather Service Cooperative Observer Network.

Current data

The 4-inch bare soil sensors were installed at the eight selected stations in the spring of 2000 and hourly data became available on 1 August 2000. The hourly data from 1 August through 18 November were used to develop a relationship between the 10:00 a.m. and daily maximum and minimum 4-inch bare soil and 4-inch sod temperatures. Separate relationships were developed for each of the eight stations, and an overall relationship using all the data from the eight stations was formulated.

A plot of the data (Figure 2) shows that a linear relationship is the appropriate fit for these data. The slope of the data shows that there is not a 1:1 relationship between the bare soil and sod temperatures. The slope is such that under warm soil conditions, the sod temperature is cooler than the bare soil, but warmer than the bare soil when temperatures are cool, agreeing with Baker (1965). The point where the sod and bare soil temperature are equal is approximately 72 °F.

The standard error for the eight stations ranges from ±1.18 °F at Orr in western Illinois to ±2.65 °F at Stelle in east-central Illinois (Table 3). When the data from the eight stations are combined into a single linear relationship, the standard error increases to ±3.05 °F. The larger standard error is due to the varying conditions across the state. Conditions that vary include the length to which the grass is cut, thus changing the amount of shading by the plants. Stelle is a station with tall grass over the sod temperature sensor. Orr, on the other hand, has very closely cropped grass, with minimal shading of the soil containing the sod temperature sensor.

The stations in Table 3, Table 4, and Table 5 are ordered from south to north except for Kilbourne and Orr, which are ordered east to west from Bondville. This ordering shows that the intercept tends to decrease from south to north and with decreasing vegetative cover. Brownstown and Bondville as well as Stelle tend to have thicker vegetation in the sod area around the station and tend to resemble grazed pasture grasses. The other stations are more closely cropped and resemble lawns.

The slope computed using all station data for the 10:00 a.m. soil temperature relationship is equal to the slope at Orr and DeKalb. Using the equation for all stations results in the greatest error in the temperature measurement at Dixon Springs. Fortunately, for the purposes of this project, Dixon Springs is located in the area where fall application of N is not recommended.

The maximum and minimum temperature relationships respond similarly to the 10:00 a.m. relationship. Both the maximum (Table 4) and minimum (Table 5) temperature relationship standard errors of the estimate are slightly greater than standard error for the 10:00 a.m. relationship. Maps of the daily maximum and minimum 4-inch bare soil temperatures will therefore have a greater error than the 10:00 a.m. map. Thus, the maximum and minimum maps should only be used as general estimates of the daily extreme soil temperatures across the state.

Because the slope of the 10:00 a.m. relationship is greater than one, the greatest deviations of the predicted from the actual temperatures occur at the highest and lowest temperatures. By sorting all the data in ascending order, an estimate of the error of the predicted temperature can be determined for different classes of bare soil temperature. Although the soil temperature data ranged from 38 ºF to 86 ºF, the temperature range of most interest to this project is from 45 ºF to 65 ºF. This brackets the thresholds of 50 ºF and 60 ºF for nitrogen application. The range of 45 to 55 ºF was used to represent the error around the 50 ºF threshold and the range 55 to 65 ºF for the 60 ºF threshold. The root mean square error at all eight stations for the 50 ºF threshold was ±2.32 ºF with a mean error of 0.72 ºF. Thus, for temperatures around 50 ºF, there is a tendency to overestimate the actual temperature. The root mean square error for the 60 ºF threshold was ±2.40 ºF with a mean error of -0.89 ºF, a slight underestimate. These mean errors would result in start of application of nitrogen with a nitrification inhibitor when soil temperatures are warmer than they should be. However, for application without a nitrification inhibitor, the soils would tend to be cooler than the threshold, an error on the conservative side.

Historical data

Concurrent 4-inch bare soil and 4-inch sod temperatures were not readily available for Illinois stations. However, such observations were available from the National Weather Service Cooperative Weather Station located at the Purdue Agronomy Farm near West Lafayette, IN. Data from this station were obtained for the period of 1 January 1994 through 31 December 2000. Only the daily maximum and minimum temperatures were available; thus, the equation developed from these data cannot be used for this study. However, the regression coefficients for the daily maximum and minimum soil temperatures can be used to verify the robustness of the equations developed for the limited Illinois data.

The West Lafayette daily maximum equation had an intercept of -2.53±0.50 °F and a slope of 1.11±0.01. This equation used all 365 days of data each year. If only the days from 1 August through 28 November were used, the intercept was -4.43±1.41 °F and the slope 1.13±0.02. The intercepts and slopes of these two equations are not significantly different from each other. However, the West Lafayette daily maximum equation coefficients are significantly different from the Illinois coefficients. The differences in the West Lafayette daily minimum temperature equations show the same response.

From this analysis using daily maximum and minimum 4-inch bare soil and sod temperature, it is difficult to say definitely that the Illinois relationship will not change with more data. Based on the response of the West Lafayette equations on the full and partial year data, it is probable that the size of the intercept will change in the Illinois equations as winter, spring, and summer data are added to the autumn data. The magnitudes of the standard errors of the West Lafayette equation are approximately the same as those for the Illinois reduction in the Illinois standard error is unlikely. Thus, the equations developed from the Illinois data are, at this time, the best available. However, they should be revisited as more data are collected.

2001 Bare Soil Temperature

The change of bare soil temperature patterns for the autumn of 2001 shows that the temperature dropped below 60 °F in the northern two-thirds of the state during the week of 26 September (Figure 3), rebounded above 60 °F during the week of 3 October, and returned to below 60 °F during the following week (10 October). The 4-inch bare soil temperature fell below 50 °F for the first time during the week of 17 October, rebounded above 50 °F the following week, and fell below 50 °F during the last week of October.

This course of soil temperatures is typical for any given year. Based on the very warm autumn that has occurred in Illinois this year, it is surprising that the soil temperatures have remained below 50 °F since the end of October.

A major concern about the use of this product is the tendency to assume that the first time the soil temperature drops below a threshold, it is okay to apply nitrogen. To evaluate how this might affect nitrification, the measured 4-inch bare soil temperature data was plotted for Bondville (Figure 4). Plots of other stations show the same pattern as Bondville with similar temperature each day; therefore, only Bondville is used as an example.

The first day that the 4-inch bare soil temperature was below 60 °F was 25 September, and the first day the bare soil temperature was below 50 °F was 6 October. Between 25 September and 6 October, the bare soil temperature at 10:00 a.m. CST was above 60 °F three times. The degree-days above 60 totaled 4.8. Between 6 October and 28 November, the last day in this analysis, there were 16 days on which the 10:00 a.m. bare soil temperature was above 50°F. During these 16 days, which represents more than one third of the days during the period, a total of 88.3 degree-days above 50°F were accumulated.

Summary and Conclusions

The ISWS has a network of 18 climate-monitoring stations that has measured, in addition to above-ground weather variables, hourly soil temperature under a sod surface since the late 1980s. The temperatures under sod provide a close approximation to soil temperatures under field soils covered by residue. However, planting and fall nitrogen fertilizer application decisions are made based on bare soil temperatures. While the climate data have been gathered for a long period, the data were not readily available to agricultural producers or the general public.

In this project, a website was created to disseminate the daily and hourly summaries of the climate data that have been and continue to be collected. Maps and hourly data are available for the previous day by 3:00 a.m. each day. In addition to the website, soil temperature sensors were installed at eight of the climatemonitoring stations to measure 4-inch bare soil temperature along with the 4-inch temperature under a sod surface. These data were used to create a relationship to convert the sod temperature measured at 4-inches to equivalent 4-inch bare soil temperature.

Maps were generated and made available online at http://www.sws.uiuc.edu/warm/soiltemp.asp beginning the second week of October. The maps showed the previous day's 10 a.m. 4-inch soil temperature at the eight climate monitoring stations where the 4-inch bare soil temperature was being measured. Other weather variables measured at the stations were made available on the website in early January 2002.

The maps allow users to see the relative pattern of the different weather variables across Illinois. Hourly weather may also be viewed in either tabular or graphic form for each station individually. These tables will eventually allow access to the hourly data for the complete climate record of each station.

Quality control of the data collected is accomplished each day before release of the data by an automated computer system developed as part of this project. Potential errors or data problems identified by the computer program are checked manually to ensure capture of all questionable data. This combined method also ensures that equipment malfunctions are detected and repaired in a timely manner.

Tables and Figures

Table 1. Locations and starting dates of automated soil temperature data collections records at the ICN weather stations

Table 2. Soil chatacteristics at the Illinois Cliimate Network sites

Table 3. Relationship between 10:00 a.m. 4-inch bare soil and 4-inch sod temperature at the eight ICN stations with both bare soil and sod temperature readings

Table 4. Relationship between the daily maximum 4-inch bare soil and 4-inch sod temperature at the eight ICN stations with both bare soil and sod temperature readings

Table 5. Relationship between daily minimum 4-inch bare soil and 4-inch sod temperature at the eight ICN stations with both bare soil and sod temperature readings

Figure 1. Date of last temperature above either 60°F or 50°F in the autumn of the year

Figure 2. Relationship between 10:00 a.m. 4-inch bare soil and 4-inch sod temperature

Figure 3. 4-inch bare soil temperatre at 10:00 a.m. CST on various dates

Figure 4. Bondville 4-inch bare soil temperature from 1 August to 28 November 2001

Footnotes and References

¹ Steven E. Hollinger is a Senior Professional Scientist and Robert W. Scott is an Associate Professional Scientist, Illinois State Water Survey.

Baker, D. G. 1965. Factors affecting soil temperature. Minn. Farm Home Sci. 22:11-13.

Hoeft, R. G. 1998. Soil testing and fertility. In. Illinois Agronomy Handbook 1999-2000, Univ. of Illinois Extension Circular 1360, Champaign, IL. pp. 247.

Hollinger, S.E., B.C. Reinke, and R.A. Peppler. 1994. Illinois Climate Network: Site Descriptions, Instrumentation, and Data Management. ISWS Circular 178. Illinois State Water Survey, Champaign, IL. pp. 62.

Rodskjer, N., M. Tuvesson. 1976. Observations on soil temperature under short grass cover and in variously managed bare soils at Ultuna, Sweden, 1968-1972. Swedish J. of Agric. Res. 6: 243-246.

Back to 2002 Index

Department of Crop Sciences University of Illinois Extension College of Agricultural, Consumer, and Environmental Sciences University of Illinois at Urbana-Champaign Illinois Fertilizer & Chemical Association Illinois Department of Agriculture

Design by: Crop Sciences Computer and Web Support Group Copyright © 2005 University of Illinois at Urbana-Champaign Email site problems to the webmaster