1	The Ohio State University Soil Blend Model: Verification and Application Terry J. Logan, Ph.D. Logan Environmental Inc.
2	Introduction Soils for various uses can be manufactured from mineral and organic by-products. Specific uses include general landscaping, DOT spec soil, sports turf, greenhouse mixes, container media. Examples of mineral by-products include fly ash, foundry sand, steel slag, dredge spoil. Examples of organic by-products include EQ biosolids, manure compost, yard waste compost, pulp and paper sludge. This paper describes the OSU Soil Blend Model and it verification.
3	The OSU Soil Blend Model The Model uses the concept of reverse engineering to set blending ratios for the mineral and organic by-products. The Model uses a combination of soil physical and chemical properties, values for which are selected by soils experts or are specified (e.g., DOT topsoil spec). Samples of the by-products to be used are analyzed for the soil properties to be specified. Analytical data is input to the Model and least squares optimization is used to set the mixing ratios.
4	The OSU Soil Blend Model The Model can be constrained to optimize use of one or more of the by-products (e.g., it would be more profitable to use a by-product with a tip fee rather than one that had to be purchased). The Model is run in Excel with input, output and optimization modules. The Model has been used in a research mode for the last 10 years but has not been rigorously verified. The present verification was funded by a grant from USEPA Region V to the Great Lakes By-Products Management Association (GLBMA).
5	Model Verification Six by-products were selected for the verification study: Mineral - class F fly ash, steel foundry sand, steel slag Organic - biosolids compost, steer manure compost, yard waste compost Four soil materials were selected for manufacture: Landscape soil Sports turf soil Greenhouse mix Container mix
6	Soil Chemical and Physical Properties Measured Chemical - soluble salts, organic matter, total N, P, K. Physical - bulk density, total porosity, available water holding capacity. A range of acceptable values for these parameters were selected for each of the four soil mixes.
7	Chemical and Physical Parameter Specifications
8	Chemical and Physical Analysis of the By-Products
9	Volumetric Soil Blend Ratios
10	Comparison Between Soil Specs and Measured Soil Property Values
11	Conclusions Most measured values deviated from Model-predicted values by less than 50 %. Two outliers were soluble salts in the sports turf soil and total Kjeldahl-N in the greenhouse mix (the measured N value is probably analytical error). There appeared to be positive bias in bulk density and negative bias in total K. Specification values for total K were found to be set too low. All measured values were in the range specified for the mixes.
12	Conclusions The Model was accurate at the levels of precision found in soil blending. Variability includes by-product inhomogeniety, mixing errors, and inability of the Model to optimize all parameters. The Model is limited in precision when by-products are extreme in their soil properties. All mixes made had excellent visual characteristics, and preliminary pot studies with annuals, perennials and grass gave growth similar to commercial bagged topsoil. Use of the Model requires substantial knowledge of soil science and agronomy.