Of the five combustion products measured in heavy-duty vehicle emission testing (carbon monoxide [CO], carbon dioxide [CO2], hydrocarbons [HC], oxides of nitrogen [NOx], and particulate matter [PM]), PM is the only emission that eludes real-time measurement. To date, filter sampling of PM over the duration of a test is the only accepted means of PM measurement for emission standards certification. However, this provides no means for the researcher to ascribe PM generation to any specific portion of a transient test, or to correlate PM production with engine load, speed, and change of speed. Numerous attempts have been made to determine. real-time PM production. Continuous opacity is one method, but research has shown little correlation between opacity and PM. Gravimetric PM data have been measured for a range of steady state operating conditions for a particular engine to form a matrix of data used to predict real-time PM production in transient operation. This is limited to the specific engine and the correlation is marginal. The tapered element oscillating microbalance (TEOM) shows the most promise in determining real-time continuous PM. Yet observations of TEOM data suggest moisture from combustion biases TEOM data through adsorption and desorption from the TEOM filter. Moisture desorption from the TEOM filter is evident from negative TEOM mass rate data during specific segments of transient tests. The purpose of this study is to develop a model that predicts the moisture effect on the TEOM filter. Compensation for the moisture effect in TEOM data will yield a more accurate trace of legitimate PM. For this study, used and new TEOM filters were placed in an environmental chamber and mass readings were taken at five different relative humidity settings. Results showed a strong linear correlation between moisture mass on the filter and ambient moisture content in the air. Results also showed used filters adsorbed more moisture than new, suggesting filter moisture retention is a partial function of accumulated PM on the TEOM filter. A second method used to evaluate the effects of moisture on a TEOM filter was to inject steam into the exhaust pipe just after the engine exhaust manifold. The steam was injected at a known mass flow rate as a step-wise function while the engine was run at a steady state. TEOM data were collected during these tests and a model was developed to predict the moisture adsorption and desorption on a mass rate basis. The resulting model had a strong linear correlation with the TEOM data collected during steam injection. A final approach was to apply the model to a series of transient engine tests to evaluate its performance. Moisture mass rates from engine exhaust were determined using CO2 data and a moisture relation C/CO2 derived from a lean-burn combustion equation. In both the steam injection tests and the transient tests, variables in the model were determined by iteration using a forward derivative Newtonian solver. Model variable values determined from both methods were similar. Subtraction of predicted TEOM moisture mass rate from original TEOM data over the course of a transient test yielded significant reduction in negative TEOM mass rate data. Results also suggest PM mass rates during heavy engine loading to be less than reported by original TEOM data. These results affect the attribution of PM production to transient engine operation, as well as the use of TEOM data for inventory modeling.