Improved compost mix great on the pad,
but slow to deliver in the field

The good things about compost are hard to measure in a single crop year on variable soils, even with careful scientific observation.

By Christine Ziegler Ulsh, TRI research editor

Field variation weakens statistical significance

To explain this concept, we offer the following example. Each soil N (nitrogen) data point that we report for the standard dairy compost treatment is drawn from an average of four soil N measurements, each taken from a different field replication group.

The statistical power (significance) of the averaged data point we report is strongly influenced by the numerical variation among the four measurements from which it’s averaged: If the range of the four N measurements is large (say 5, 10, 15, and 20, with an average of 12.5), then the statistical significance of the averaged N data point is very low, because 12.5 doesn’t very well represent the readings of 5 or 20. Conversely, if the four N measurements are very similar (say 11, 12, 13, and 14, still averaging 12.5), then the significance of the averaged N data point is very high (12.5 is quite representative of 11 and 14).

Therefore, if the basic soil N level is 5 in Replication (Rep) A and 20 in Rep B, then the N data point averaged from measurements taken from standard dairy compost treatment plots in these two replications (12.5, in this case) is not statistically representative of the two readings.

Even if the compost treatment manages to raise soil N levels by a significant amount, such as 10 points in each plot (making readings of 15 for Rep A and 30 for Rep B), the average N data point of 22.5 is still not statistically representative of either 15 or 30. Thus, even a significant change can be “hidden” by variations in the field and a low number of replications.

Posted September 14, 2007: Composting is an effective, well-known but vastly under-utilized means of managing crop and livestock wastes and building soil fertility. Proper composting converts crop residue and animal manure into a soil amendment that provides slow-release, balanced crop nutrients and greatly improves soil structure by increasing soil organic matter. The Rodale Institute research staff is working to find better ways of producing and using compost on-farm, so more farmers can apply compost on more acres with better results.

In April of 2006, we reported information drawn from the first year of our Pennsylvania Department of Environmental Protection-funded compost research project. After composting manure alone, manure with leaves, and manure with leaves and amendments, we found compost mixes go a long way to reduce nitrogen, phosphorus, and pathogen losses during the composting process, especially when the pile is amended with a mixture of clay, calcium and humic acid. (See the New Farm article Good compost made better.)

Starting that month, we began field testing the compost we had produced. Given our success in reducing nutrient losses from the compost piles, we were excited to be able to test these composts and aged manures along side raw manures and fertilizer in the field. We wanted to see if our compost mixes would support respectable corn yields and help retain soil nutrients as well as they had in the pile.

One-year results are not so clear

After a year of corn production, the short answer concerning the impact of these composts on soil quality and corn yield is that we still don’t know. While we did find a few significant differences in soil-nutrient levels and nutrient leaching rates over the course of the year, they were not clear enough to draw any strong conclusions. At the same time, we found conventional fertilizer resulted in higher corn yields than the amended dairy compost, but differences among all the other treatments weren’t significant. We did learn that uncovered winter storage leached nutrients from our finished poultry composts. However, the biggest lesson we’ve taken from our data this field season is that, in order to get the picture of what compost does for the soil or a crop, you’ve got to use it for at least three to five years.

The more complete story began when we tested all our composts, aged manures and raw manures in order to determine their nutrient content and calculate spreading rates. Sadly, the winter had not been kind to our finished poultry compost. When we pulled the finished poultry compost from the composting pads in October 2005, the N-P-K ratios of the treatments were 0.6-1-0.7 for the aged manure, 1-1-0.7 for the standard poultry compost, and 1.6-1-1 for the amended poultry compost. (The ideal N-P-K ratio for most crop fertilizers is 2-1-2.). However, when we went to apply the composts to the field in April, the N-P-K ratios of the treatments were as follows:

TREATMENT
N
P
K
fresh poultry manure
0.4
1
0.3
aged poultry manure
0.6
1
0.5
standard poultry compost
1
1
0.5
ammended poultry compost
1.1
1
0.5
fresh dairy manure
1.8
1
2.7
aged dairy manure
1.3
1
2.2
standard dairy compost
2.6
1
2.3
amended dairy compost
2.4
1
1.5
conventional fertilizer
2
1
2
TARGET
2
1
1

While the ratios of the standard and amended dairy composts were excellent, the poultry composts had clearly lost both N and P in piles over the winter. We had no way to measure water leached from the pile, so we cannot say whether more of this N was lost to volatilization or leaching, but the P was lost through runoff and leaching, as it does not volatilize. However, in either case, we would likely have been able to conserve both N and P if we had covered the piles. Therefore, we recommend that, if compost cannot be used as soon as it is finished, it should be covered (either by roof or tarp) to prevent these nutrient losses.

As we prepared to apply these manures and composts to the field, we needed to estimate the availability of the N in each treatment in order to calculate an application rate that would supply 150 pound of N per acre for our corn crop. As farmers know, these N availability estimates are, at best, good guesses; once applied to the field, actual N availability varies due to a number of factors including weather, tillage, soil biota and soil-carbon levels, to name a few. We estimated the raw and aged manures would make about 50 percent of their N content available to the corn crop, while the standard composts would release 40 percent of their N, and the amended composts would free up 30 percent of their N, more or less. Thus, given the varied amount of N in our composts and its varied availability, our application rates for each treatment were as follows:

TREATMENT
TONS PER ACRE
fresh poultry manure
11.4
aged poultry manure
10.7
standard poultry compost
15.6
ammended poultry compost
10.7
fresh dairy manure
17.9
aged dairy manure
18.1
standard dairy compost
23.5
amended dairy compost
26.4


The compost was applied on May 9, 10, and 11, and it was plowed under, along with the rye cover crop, on May 11 (Figure 1). We laid out the research field so the nine fertilization treatments appeared once in each of four replication groups, totaling 36 plots. After allowing time for the rye to begin decomposing, field corn (Blue River 68F32) was planted May 23 in all 36 plots at a rate of 30,000 seeds/acre (with starter fertilizer included for the conventionally fertilized plots).

Prior to the compost application and corn planting, we took soil samples from each of the 36 field plots to assess baseline levels of C, N, P, K, pH, other macro and micro-nutrients, and organic matter. We found there was a significant pre-treatment difference in organic matter between the plots on which we applied the standard dairy compost and the amended dairy compost. But more importantly, we also found significant differences among the four replication groups for almost every soil nutrient and parameter we measured, indicating each section of the field varied remarkably. This finding concerned us, statistically speaking, because differences among replications can easily obscure any significant differences created by our treatments. (For details, see the “Field variations” sidebar.)

Of course, because field variation is an inherent part of agronomic field research, this discovery did not deter us from our soil sampling plan. We gathered subsequent soil samples in mid-June, early December (after corn harvest), and in May of 2007 to see if and how soil nutrient levels varied over that time period and among the treatments.

At the same time, we collected water leached from the soil below the root layer to measure the amounts of nutrients lost from the field, using devices called intact-core lysimeters. (Click here for details.) Beginning in April 2006, water was collected eight times through the ensuing year (ending in May of 2007) and analyzed for nitrate, ammonium, ortho-phosphate, electrical conductivity (EC), pH, and total dissolved solids (TDS).

Throughout the season, we also tracked corn-plant nitrogen levels by measuring ear-leaf chlorophyll levels with a hand-held chlorophyll fluorescence absorbance meter (a Minolta SPAD Meter – Specialty Products Agricultural Division) and via season-end corn stalk nitrogen sampling. Then, finally, at season's end, we collected harvest yield data and had the corn grain analyzed for nutrient content.

So what did we learn from all this work? In a nutshell, we found:

  1. None of the composts provided enough N at the levels we applied them to support yields that were comparable to the chemical fertilizer, at least not in the first season of use.
  2. The composts gave yields comparable to those produced by the aged and raw manures with the exception of the amended dairy compost. However, these statistics are likely diluted by the fact that the replication groups also showed significant yield differences.
  3. SPAD chlorophyll readings appear to be quite accurate at predicting corn yields, much more so than corn stalk nitrogen sampling.
  4. Raw poultry manure leached more nitrate and dissolved more solids into the soil water than any of the other treatments (significantly more than aged poultry manure).
  5. There were some sharp variations in the leaching of ammonium and phosphorus, but they were not significant and appeared to be due to variations in the field.
  6. The poultry manure composts (both standard and amended) significantly improved soil-carbon levels when compared to the chemical fertilizer (soil carbon is a key element of soil organic matter). In fact, the effect of these composts was enough to eliminate the significant soil-carbon differences among the field replication groups.
  7. The raw poultry manure generated the highest levels of soil phosphorus, potassium, and CEC (significantly higher than the chemical fertilizer in each case), but these results are questionable due to lingering significant differences among the replication groups.
  8. All of the above findings would either a) gain greater statistical significance and relevance, or b) change in potentially significant ways if we continued the study for three to five field seasons.

Points 1, 2, and 3 are well illustrated in Figures 4 and 5.

 

As you can see, the conventional fertilizer produced the highest yields and amended dairy compost gave the lowest. The fresh and aged poultry manure, as well as the aged dairy manure, also gave statistically higher yields than the amended dairy compost, but all the other treatments were not statistically different, though there are some clear numerical differences. What’s more, the chlorophyll data mirror the yield data almost exactly with greater statistical significance, since the SPAD sampling yielded many more data points to average. This comparison suggests that SPAD chlorophyll readings may be a very good way to estimate yields before harvest, even early in the growing season (our SPAD data from September was quite similar to that which we had collected in July). However, we hope to see this comparison borne out over several more growing seasons before recommending SPAD readings to farmers as means of assessing plant nitrogen needs.

Difficult determinations

We weren’t surprised by the high nitrate leaching rate of the raw poultry manure, given the material’s high volatile N content. Unfortunately, our other leachate data were badly confounded by one field plot in the conventional fertilizer treatment that leached remarkably high amounts of ammonium, and another field plot in the amended dairy compost treatment that leached excessive amounts of phosphorus. We tried to remove these “outliers” and redo the statistics, but the remaining three data points weren’t enough to create a statistically significant average. This is a classic example of the challenges inherent in agronomic field research.

Finding four or more field plots of adequate, farm-scale size that are similar in soil characteristics and topography is a difficult (if not impossible) task in Pennsylvania, and even similar plots may still behave differently for reasons we don’t understand and can’t control. Weather, disease, or pests can also wipe out an entire year’s data. This is why multiple-year replication of field crop experiments is vital. Many years are required to compile enough clean (statistically significant) data points on which to base solid recommendations.

Our most promising results are the soil-carbon increases generated by the poultry composts. The significant carbon increases in these two treatments, and the fact that all the manure and compost treatments increased soil-carbon levels when compared to the chemical fertilizer, suggest that carbon-rich fertilizer amendments really can improve soils, even in one year. This point is borne out by The Rodale Institute’s 10-year-long Compost Utilization Trial which showed that continuous compost use significantly increases soil carbon levels and also supports comparable crop yields after three to five years of use.

Implementing nitrogen

We suspect crop-yield delays in initial years of compost use are caused by the slow-release nature of compost-based nitrogen. While mineral nitrogen dissolves easily in water and thus is quickly available for crop uptake (or leaching, if the crop isn’t growing well), compost-based nitrogen is very stable within soil aggregates and must be released through microbial action. Therefore, we recommend that farmers who want to start fertilizing with compost begin by supplementing their crops with some form of quick-release nitrogen during the first two years of compost use, with both starter fertilizer and side-dressing nitrogen during the growing season (Chilean nitrogen is an approved, though expensive, option for organic farmers). We have begun looking for ways to develop compost “nuggets” that might be used as a starter fertilizer or for side-dressing, but this work is new and no results are yet available.

Our ultimate “take-home lesson” from this trial is that any future compost research needs to last a minimum of four years: one year for initial compost production and then three years of continued compost production and field application. We would also decrease our initial estimates of N availability for the amended composts (and possibly for the standard compost) to compensate for the slow-release nature of the nitrogen and allow soil microbiota to “get up to speed” in converting that nitrogen into forms available to the crop. We might need to then revise those N-availability estimates after several years of use, but initial heavy compost applications will likely improve crop yields during the conversion of chemically fertilized fields to organic management.

We are grateful to the Pennsylvania Department of Environmental Protection for their funding and support of this project. We look forward to partnering with them in the future to study the ways compost and other agricultural practices can be used to sequester carbon in the soil.