The Intermountain West is topographically and climatically diverse and provides many different habitats to study the relationship between weather and fungal abundance and species richness. We selected 25 field sites (approximately 1 hectare in area) in and near Utah that differ in factors that influence weather such as elevation, longitude and latitude, proximity to permanent water sources, and topography. To eliminate comparing habitat extremes, all 25 field sites share a widely distributed species of shrub, sagebrush, Artemisia tridentata (Asteraceae). Community types include arid grasslands with cacti, riparian communities characterized by cottonwoods, and subalpine areas with tall forbs and aspen stands.
The Intermountain West consists of two biogeographic provinces, the Great Basin and the Colorado Plateau, that differ in precipitation patterns and drainage . While both the Great Basin and the Colorado Plateau share weathers with cool wet winters and relatively dry summers, they differ in that the Colorado Plateau has wetter summers and drier winters than the Great Basin . There is also a trend for sites to have lower total annual rainfall towards the south and west.
Fungi were collected in the air and on the vegetation (see below) and were grown on plates containing growth medium. This method quantifies a portion of the viable propagules that are able to germinate and form visible colonies in the time frame observed unlike visual counts of spores using a microscope. Four types of media containing 2% granulated agar (Difco, Detroit, MI) were used to grow fungi: (1) potato dextrose agar , (2) potato sucrose agar (similar to potato dextrose agar except sucrose is used instead of dextrose; , (3) casein agar (for 1 L, 5 g casein, 0.5 g NaCl, 20 g sucrose, and 20 g agar), and (4) sagebrush-leaf agar. To prepare the sagebrush-leaf agar, we removed surface secondary metabolites on 80 g of fresh A. tridentata leaves by washing them in 400 ml of chloroform for three minutes and washing again in 200 ml of chloroform for 1 minute. Leaves were placed under a hood for one day to remove chloroform and then dried at 70°C to constant weight (1 to 2 days). We homogenized dried leaves in 900 ml of water and then filtered the leaf extract (Whatman #1). We combined 500 ml of leaf extract, 250 ml of H2O, 20 g of agar, and 250 ml of Czapek's minerals ; 1 L of Czapek's aqueous solution, contained 2 g NaNO3, 1 g KH2PO4, 0.5 g MgSO4 · 7H2O, 0.5 g KCl, 0.01 g FeSO4, and 1 ml of an aqueous solution of 1% ZnSO4 and 0.5% CuSO4). All four media contained ampicillin (60 μg/ ml) to eliminate contamination by bacteria. We used potato-based media because they are commonly used to grow fungal pathogens. We used the leaves of A. tridentata in a medium (sagebrush-leaf agar) because this species is common to all habitats.
During September of 1998, we measured fungal abundance and richness in each habitat by culturing fungi found on the vegetation in 21 communities. Because some plants contain secondary metabolites on the leaf surface that inhibit fungal growth, we cultured fungi collected only from leaf surfaces of the species that were not aromatic when their leaf surfaces were rubbed. Using sterilized scissors, we collected one cm2 of leaf material from each of seven haphazardly chosen plants (from three to five species) and combined all seven in a single sterile vial. This was repeated using the same plants. Leaf samples were collected in the late afternoon and stored at 4°C for up to 7 days before culturing. The relationship of diversity and abundance of culturable fungi with storage time was not determined. To remove fungi from leaf surfaces, we soaked and vortexed leaf samples for 15 seconds in 7 ml of a sterile aqueous solution containing a 0.5% sodium chloride, 2.0% casein, and 4.0% sucrose. We spread 250 μl of the inoculated solution onto one 10-cm Petri plates each of potato sucrose agar, casein agar, and sagebrush-leaf agar. Since we collected two vegetation samples per site, there were two plates of each of the three media, or six plates per site.
We incubated the inoculated plates at room temperature (22°C) for 4 days before counting the number of colonies per plate as a measure of fungal abundance. We found little change in the number of colonies after 4 days, presumably due to the proliferation of the aggressively growing fungi. We classified the colonies under a disecting microscope (10 to 50×) into morphotypes based upon colony morphology (effuse, raised, domed or umbonate), color, transparency, and shape (irregular, crenated, or rhiziod). We scored fungal richness as the total number of morphotypes per plate. We averaged fungal measures from two vegetation samples for each of the three media. We did not attempt to identify species or to compare morphotypes across plates. Although the different media probably favor different fungal species, our goal was not to measure habitat differences of specific fungi, but to estimate the overall fungal abundance and species richness of the habitats. Furthermore, the fungal measures likely represent the abundance and richness of mesophiles, because the temperature (22°C) used to incubate inoculi falls within the temperature optima of mesophilic fungi.
During September of 1998, we collected air-borne fungi at 17 sites by vertically swiping three agar plates (one each of potato sucrose agar, casein agar, and sagebrush-leaf agar) ten times in the air. During October of 1999, we collected air-borne fungi at 22 sites by swiping 10 potato dextrose agar plates vertically 10 times in the air. At each site the replicate plates were collected at 10 haphazardly chosen locations to obtain an average measure of fungal abundance and richness of the habitat. Samples were collected in the late afternoon. We incubated inoculated plates in a cooler at ambient temperature (approximately 22°C) for 4 days and examined plates in the field. Fungal abundance and richness were scored as above.
Direct weather measurements
We recorded the average temperature and relative humidity every 30 minutes at 19 of the 25 study sites with NOMAD dataloggers (OM-NOMAD-RH-32, Omega Engineering, Inc; Stamford, CT) protected in a weatherproof casing. For all sites, we recorded the weather for a complete growing season, mid-March of 1999 to mid-October of 1999. To compare annual differences in weather, we also recorded weather from June of 1998 to mid-October of 1998 at seven sites. To investigate differences within a habitat, two sites were equipped with two dataloggers each. Using these data, we calculated weather variables that are commonly reported in the literature (see below).
We calculated the following temperature parameters in °C for the entire growing season: average, standard deviation (of the 30-minute readings), range, maximum, and minimum. We also calculated a "length of growing season" parameter equal to the maximum number of consecutive days without freezing temperatures.
The growth and survival of a fungal spore germinating on a leaf depends on the amount of moisture available in the air and the length of time the leaf surface is wet . Dew formation can be critical for spore germination of many plant pathogens, and may be the best indicator of the proportion of time a habitat has weather favorable to fungi. Although the duration of leaf wetness is frequently measured in agricultural studies , it is not practical for many ecological studies. However, relative humidity of the air is a good measure of moisture availability . The relative humidity (RH) where RH = (e
) × 100% (the units of e
are kPa) is the ratio of water vapor pressure in the air, e
, to the maximum amount of water vapor that air can retain at a given temperature, the saturation vapor pressure, (e
) where e
= 0.61365e17.502T/(240.97+T) (the units of e
are in kPa and T is the Celsius temperature; Buck 1981). Using the relative humidity records from the dataloggers, we calculated the following relative humidity parameters for the entire growing season: average and the standard deviation (of the 30-minute readings), average of daily maximums, and the seasonal range, maximum, and minimum.
Vapor pressure deficit
The vapor pressure deficit (VPD) where VPD = e
× RH/100) (the units of e
are in kPa) is the maximum amount of water vapor that can be retained in air at a given temperature minus the actual water vapor pressure in air . This measure reflects the desiccating capacity of the air. Using the 30-minute readings of relative humidity and temperature from the dataloggers, we calculated the average VPD and the standard deviation for the entire growing season.
Proportion of time a habitat has weather favorable or unfavorable to fungi
We calculated weather variables that reflected the proportion of time a habitat is favorable or unfavorable for fungal growth. We calculated the proportion of time (based on the 30-minute averages over the growing season) that the temperature was between 4 and 40°C  and also had a relative humidity greater than 94%. Three other indices were calculated as follows: the proportion of days when the temperature remained above 4°C and below 40°C as well as having a maximum relative humidity that was (i) greater than 90%, (ii) greater than 75%, or (iii) less than 50%. A broad temperature range (4 to 40°C) was used to represent the range of temperatures likely to support most mesophilic fungi [2, 12, 15, 18]. Because the growth of mesophilic fungi can occur between 0 and 50°C with optima from 15 to 40°C, a temperature of 4°C may represent a conservative minimum at which mesophilic fungi grow and a temperature of 40°C may represent a conservative maximum at which mesophilic fungi grow.
Duration of periods with a favorable or unfavorable weather
We calculated weather variables that reflected the duration of consecutive favorable or unfavorable days in a habitat. We estimated the average number of consecutive days during the growing season when the daily temperature remained between 4 and 40°C as well as having a maximum diurnal relative humidity that was (i) greater than 90%, (ii) greater than 75%, or (iii) less than 50%. The maximum number of consecutive days also was calculated for the same three parameters.
Climate estimated by nearest weather station
We estimated the climate of each site using the monthly average precipitation and temperature collected by weather stations located 4 to 30 km from the study sites. For three of the high elevation sites (Francis, Racetrack, and Strawberry in Northern Utah), we used records from Alta, Utah, a site with similar altitude and location (within 100 km of the three sites). Climate data were obtained from the Western Regional Weather Center and elsewhere . Climate records covered 21 to 58 years, depending on the site.
We scored two other site-specific factors, the presence of nearby water sources and vegetation types indicative of moisture [34–36]. For the presence of nearby water sources, we gave a score of zero for none, 0.5 for ephemeral water sources such as washes, and a score of 1.0 for habitats with permanent water sources such as streams and lakes. We listed the five most common plant species at each site (determined visually) and scored sites for vegetation types indicative of moisture. Plant species characteristic of dry habitats such as cacti and yucca were given a zero. Each plant species indicative of intermediate levels of moisture, such as juniper, pinyon pine, scrub oak, and rabbit brush were given a score of 0.5. We gave a score of 1.0 for each plant species characteristic of mesic habitats such as riparian vegetation, tall forbs, aspen, and spruce. Therefore, the highest vegetation score a site could have is 5.0.
We also measured soil properties that might be indicators of habitat moisture. To control for variation in soil nutrients due to seasonal fluctuations, we collected soil from 22 sites within a month during the spring of 1999. Because overtopping vegetation can affect the nutrients in the soil below, we collected samples away from overtopping vegetation . We discarded the top 5 cm of soil and collected 2 to 3 kg of soil up to a depth of 20 cm. We dried soil samples for 48 hours or until constant weight at 105°C. For particle size analyses, we removed rocks before sieving . We determined the percent silt and clay using a hydrometer. We determined soil pH using 15 g of ground soil in 30 ml of distilled water. To determine percent organic carbon and nitrogen of soil by mass spectroscopy, we first removed the inorganic carbon by bringing soil samples to a pH between 4 and 5 with 5N hydrochloric acid. Soil samples were then flash combusted on an elemental analyzer (Carlo Erba, Milan, Italy) coupled to an isotope ratio mass spectrometer (Finnigan MAT, Bremen, Germany) operated in continuous flow mode.
We performed univariate and multivariate statistical analyses in SYSTAT, Version 7  and JMP (SAS Institute Inc., Cary, NC). For differences in fungal abundance and richness among media types and sites, we applied Kruskal-Wallis nonparametric analysis. Using simple linear regression, we analyzed the natural log of fungal abundance and richness as a function of the previously mentioned weather variables. For periods when the dataloggers malfunctioned and did not record temperature and humidity, we estimated missing values in SYSTAT by their correlation with data from other seasons at the same sites. For example, spring data were missing from the Canyonlands field site, so we used summer and fall correlations of relative humidity and temperature across all sites to estimate the missing value for Canyonlands. There were three sites with missing values for spring, three for summer, and none for fall. To determine whether two weather variables could explain more than a single variable, we applied stepwise multiple regression analysis.