Microbial growth in indoor spaces has been identified as the main moisturerelated cause of Indoor Air Quality (IAQ) problems. Although studies in the past several decades have been conducted to identify ways to assess, mitigate, and predict mold problems in buildings, quantified mold growth risk assessment is not yet available. In many cases, mold occurrence involves local, situational, and sometimes idiosyncratic aspects of a building during its operation. These unexpected behaviors cannot be captured by current deterministic performance evaluation methods. Hence, this study has attempted to develop a probabilistic performance indicator for mold growth risk by treating mold as a risk and a limit state phenomenon. This new approach requires a reliable aggregation method to arrive at quantified mold growth risk and the extension of standard simulation capacity to account for additional mechanisms of the mold phenomenon. It also implicates uncertainty in building parameters, including natural variation of hygrothermal properties in building materials, deviation between "asdesigned" values, and the actual "I n-use" values of the parameters.

In this study, the mold germination stage is considered a limiting criterion for risk, realized by using the mold germination graph method based on local environmental conditions calculated from hygrothermal models and a standard mold germination graph. This method keeps track of the environmental conditions (i.e., surface temperature, relative humidity) at previous time steps so that the effect of the fluctuating conditions can be considered. A comparative study using the mold limitation curves in ESP-r showed that the mold germination graph method could provide quantitative evaluation results in terms of mold gro wth risk. The current standard moisture simulation was extended to account for additional mechanisms of the mold phenomenon (e.g., thermal bridge, workmanship, infiltration), based on four identified cause categories that represent the four areas in which an extension of simulation capabilities would be needed in order to produce accurate assessments. These extensions were realized using a mix of existing simulation tools, each specialized in a particular domain of heat, air, and moisture transport. The effects of thermal bridge and potential bad workmanship in a particular building construction technology are expressed with the temperature factor. By including the temperature factor at certain building details in the mixed simulation approach, the mold growth risky days can be approximately represented at a specific tro uble spot.

The uncertainties of each building parameter are expressed as upper/lower values with a probability distribution based on available data in the literature, different models, and field measurements. The uncertainty associated with the temperature factor is investigated by establishing an empirical relationship between an idealized thermal bridge and potential bad workmanship. The quantified uncertainties of the parameters are propagated through the mixed simulation approach using the Latin Hypercube Sampling method, which is particularly suited for this purpose. The results of the uncertainty analysis provide the distribution of mold growth risk at a trouble spot in a particular building. The identification of dominant parameters that have a major influence on mold risk is performed using a parameter screening technique suggested by Morris(1991). Knowledge of these dominant parameters is vital, as they point to the areas that require special attention during design and construction in order to guarantee a mold free environment over the li fe cycle of the facility.

The application of the developed mold risk indicator (MRI) is reported in three case studies: one virtual office building in Miami and two existing buildings with mold problems in Atlanta. The results of each case study are encouraging. They indicate that a reliable distribution of mold growth risk may result from the uncertainty analysis, as the actual mold growth occurrence could be related to the established mold risk in each case. The new approach thus seems capable of explaining unexpected and non-deterministic mold growth occurrences. Moreover, it identifies the parameters that have dominant effects on the increase in mold risk. Identification of these parameters leads to recommendations and guidelines for designers and engineers to guarantee better building perfor mance.

The new performance indicator for mold risk is capable to reveal the actual mold risks going beyond the deterministic assessment. The identification of the parameters that have a major influence on mold risk may provide a long-awaited breakthrough for early control of mold risk, i.e., during design, commissioning, and A/E procurement. More calibration and validation will be necessary but the early results reported in this thesis indicate that the new mold risk indicator provides a foundation for establishing building performance criteria that will prevent mold in buildings.

• fungi: assessment, sampling and analysis