Big data from tiny flies: patterns revealed from over 42,000 phorid flies (Insecta: Diptera: Phoridae) collected over one year in Los Angeles, California, USA

Study of the 360 samples yielded 42,480 phorid flies of 99 species (Table 2), 43 (43%) of which were new to science (Hartop et al. 2015, 2016a, Hartop et al. 2016b). By far the most abundant species was Megaselia agarici, which alone accounted for roughly one quarter of all phorid flies encountered. The second most common species, M. sulphurizona, accounted for only about 12%. Five species, M. agarici, M. lombardorum, M. nigra, M. sulphurizona, and M. wiegmanae, were collected at all sites, but sixteen species were collected from only one site, followed by 11 species at two sites (Figs. 4, 11).

The species composition at sites varied widely, from 82 species at site 13 to 19 species at site 8 (Table 3). A visualization of this data can be seen in Fig. 5, where the biodiversity at each site is represented by both the size and color of the marker sphere. The percentage of Megaselia agarici ranged from about 4% of the catch (site 5) to 75% of the catch (site #21) (Figs. 6 and 7). Sites with larger numbers of specimens had more species (Fig. 8), possibly indicating the shared requirements of much of the fauna.

Pooled monthly catches strongly peaked in May, led by Megaselia agarici and the three next most abundant species (Fig. 9). Total catch for May was about 7000 specimens, 18% of the entire year’s total. Some species peaked earlier or later, and some were found almost exclusively during the (extremely mild, even by southern California standards) winter.

Some species, including M. agarici, appeared to show two strong peaks during the year. Examination of the catches from individual sites, however, shows that this pattern is an artifact of pooling data. For example, M. agarici abundance at some sites (especially 16, 18, 23) had a strong peak in May, and others (3, 8, 9, 12, 17, 19) peaked in November. Individual sites do not often show strong peaks in both months, although many do have higher numbers in those parts of the year when compared to the surrounding months. It is unknown if this type of pattern is an indication of a bivoltine biology that is expressed differently in different microhabitats/climates. The other likely explanation, supported by the patterns seen by Disney et al. (1981), is that the severe drought in California has caused lower numbers during the hot, dry, summer months. Thus, the higher numbers that should last throughout the summer instead appear as some form of dual peaks before and after the hottest part of the year. As California is still very much in drought, we cannot yet examine this phenomenon. If (when) the drought situation changes it will be interesting to see how this affects phenology data.

Besides the relatively constant domination of samples by Megaselia agarici and others, we noted significant “outbursts” of some species. Most significant were 459 M. marquezi from site 5 during April and 470 M. barberi from site 26 during March–April; unfortunately, neither has a known lifestyle.

In addition to the main BioSCAN transect, several Malaise traps have been used to do supplemental collection in select locations. Although these results are somewhat limited, they did offer some interesting data, below.

City Hall: In July and August of 2015, we erected three Malaise traps around Los Angeles City Hall. The locations were at ground level on the north side of City Hall, a canopy trap suspended in the palm trees south of the west entrance to City Hall, and on the roof of City Hall South. In total, 11 species were found from these traps, the roof trap (8 stories up) collected only 4 species. The low diversity of the samples was not surprising given the downtown (heavily urbanized) location of City Hall. What was surprising was that every sample collected from City Hall (from all three traps, all weeks) was dominated by our local Metopina species, a result not seen at any other BioSCAN site. Prescher and Büchs (1997) recorded Metopina dominating (between 50.5–72.9%) the fauna of wheat, corn and potato fields in Germany. Similarly, Froese (1992) found that 79.7% of the phorids obtained from the soils of arable fields in Germany were Metopina. Since the genus does not appear to be associated with a particular crop, it may be associated with freshly turned soil. The grounds of City Hall were recently landscaped, and an adjacent lot was being developed, offering an explanation for the dominance of this genus in these Malaise traps. A second interesting find from City Hall was that in late August the trap on the north side at ground level caught a single specimen of Megaselia seticauda Borgmeier. This species, known to be a pest of corn, was not found elsewhere in our BioSCAN sampling. We have no explanation for this oddity, but here include it to show the value of supplemental collection – you never know what one additional trap, or week of collection, will reveal!

A species accumulation curve for the pooled 360 samples is given in Fig. 10. Based on various estimators, the entire fauna is almost fully sampled, and should total between 101 and 106 species (Table 4).

Individual traps collected from 105 phorid specimens (site 8) to 3505 phorids (trap 28, Fig. 6). The catch from trap 8 is suspect, however, because in October, 2014, a crack was found in the plastic collecting head, through which many phorids might have escaped. Similarly, sites 9 and 10 had some months with no phorid catch, which seemed to us unlikely, but we have no evidence that there were any factors negatively affecting the sampling.

Four diversity statistics are calculated by Estimates, which, aside from ranking site 13 highest and site 21 lowest, gave similar results (Table 5), although differing in some details.

Analysis of mean number of species per trap showed significant diversity within sites (Table 5), but pairwise t-tests found that only some sites differed significantly from others (Fig. 12).

The most diverse site was site #13 (Figs. 2, at bottom, and 11), which is virtually natural backyard habitat in the Verdugo Mountains. The total for this site, 82 species, is remarkably higher than that for the second richest site, #31, with 59 species. The 23 species absent from site #31, but present in site #13, are mostly rarely encountered species shared with only one other site. Most species known from only one or two sites were found at site #13, including Gymnophora talea, Aenigmatias sp., and most of the ant-parasitoid species in the genera Apocephalus, Microselia, Pseudacton, and Veruanus. The restriction of ant-parasitoids to the least disturbed sites is undoubtedly due to the presence of the Argentine ant, Linepithema humile, in the more central, urbanized part of the L.A. basin. This introduced species is found almost everywhere, especially where soil has been disturbed and irrigation exists. It outcompetes native ants, rendering most of Los Angeles free of native species, as well as their commensals, parasitoids, and predators. It is unknown whether the presence of Argentine ants is responsible for the elimination of other phorids in more urbanized areas, but it does not seem impossible. A big step towards restoring Los Angeles biodiversity would be the eradication of this synanthropic ant, which is unfortunately unlikely, given its ubiquity and apparent ecological dominance. Study and importation of phorid flies in South America known to attack Argentine ants might offer some relief for native species.

Interestingly, two cosmopolitan species of phorids, Megaselia scalaris (Fig. 11) and Conicera tibialis (the “coffin fly”) (Fig. 11) were not collected at site #13 or site #31, indicating that these species might be true synanthropes, needing to live in close association with humans. Disney (2008) reviewed the natural history of Megaselia scalaris and listed, among its larval foods, paint and shoe polish (among more normal sources of nourishment like carrion, fungi, bacteria, and other insects), hinting at the incredible adaptability of this species. Another species that appeared to have a strictly urban distribution was M. ilca (Fig. 11), a North American species. Fascinatingly, we have discovered M. ilca in samples from natural habitat in the mountains on the perimeter of the Los Angeles Basin. We are curious what set of conditions has led to the bipolar situation for this species.

The least species-rich sites were 8 and 10, although data from site 8 cannot be used confidently (as mentioned above). All diversity statistics ranked site 21 as least diverse, probably because of the overwhelming dominance by Megaselia agarici. Dominance of this species was highly negatively correlated with overall species diversity per site. Its occurrence was most highly correlated (0.63) with that of the introduced fungus feeder, Megaselia nigra. Another major fungus feeder, Megaselia halterata, had a much lower correlation (0.25), perhaps due to its dependence on the mycelium of fungi rather than the fruiting bodies (Fig. 12.)

The usual urban biodiversity paradigm is to classify species as urban avoiders, urban exploiters, and intermediate urban toleraters (e.g., McKinney 2002). For our data, 40 species were urban avoiders (Table 2), many of them found only at site #13 (the species in the top row, towards the right in Fig. 11). This list includes nearly all of the non-Megaselia genera, except for the introduced synanthrope Dohrniphora cornuta, and a few others. An example of an urban avoider’s range is Anevrina variabilis, which is apparently unable to live without populations of wild rodents in close proximity (Fig. 13).

Urban exploiters include Megaselia agarici and other common fungus feeding species, which seem to benefit greatly from the apparent increased presence of Agaricus mushrooms in urban settings. The qualifier “apparent” is used in the previous sentence because we have found no surveys of urban fungi that we can cite. Wherever Agaricus mushrooms occur, however, they can have a profound effect on the phorid fauna, as we have found over one-hundred larvae in a single fruiting body of these fungi (unpublished data).

Origins of the fauna are difficult to ascertain. We assume that any species with a joint New World-Europe distribution may have been introduced into North America, or vice versa (often these species are also found elsewhere in the world). By this criterion, at least 18 out of 99 species are potentially introduced to Los Angeles. Many species are described for the first time in this study, and consequently are not yet known from any other parts of the world. Undoubtedly, however, some of these species will be found to have originated in parts of the world other than the relatively well studied parts of Europe, as have some of the other accidentally introduced species that our study has discovered (Grimaldi et al. 2015).

Finding sites that are strictly comparable to ours is difficult. Most of the original habitat in the lower parts of the basin is either converted to residential land or highly degraded by human activity. Thus, we are forced to compare with higher elevation sites in the nearby mountains. Compared with the few samples that we have so far analyzed from more natural areas surrounding Los Angeles, our catches are strongly differentiated from those at the more natural sites. In a sample from Diamond-X Ranch in the nearby Santa Monica Mountains (34.098 N 118.699 W, 215 m), Megaselia agarici comprised only 4% (9/209) of the phorid sample, which was otherwise dominated by dozens of other Megaselia species.

Durska (1981) reported on an urban survey in Warsaw, Poland, in which three types of sites were sampled: parks, housing estates, and the town center. This study predated her more intensive work on phorid flies, and includes only 1066 specimens belonging to 36 species. Their collecting technique was also highly divergent from ours, being a type of pan trap in the crowns of trees. Nevertheless, they found that species abundance and richness decreased from parks to housing estates and again from housing estates to the town center. Their most dominant species overall was Megaselia angusta, an Old World species not found in our survey and belonging to a difficult complex of species that are prone to misidentification (Disney 1999). The highest level of domination by this species was 33.3% in the samples from parks. Other dominant species were Diplonevra nitidula, Megaselia “pulicaria” (the species after which the aforementioned complex of species that includes M. angusta is named), Megaselia rufipes, and Phalacrotophora fasciata; of these, only M. rufipes was also found in our survey at relatively low levels. From their total list, the only other species shared with our project was Dohrniphora cornuta.

Durska (2001, 2002) also studied the fauna of successional growth in (non-urban) pine forests, and presented both pooled and individual species phenology charts for a site in Poland. Overall, her charts resembled patterns seen in Los Angeles for Megaselia agarici in L.A. (Fig. 9), with a large post-autumnal rise in abundance. Other phorids in our study did not show such a prominent late surge. Also, our data had its largest peak in May, whereas the pooled data in Durska’s study peaked in October. These discrepancies may be due to the geographical position of Los Angeles being much farther south than Warsaw, the previously discussed drought in California, or a combination of the two.

The list of phorids found in Malaise trapping in the Buckingham Palace garden in England (Disney 2001) included a number of species that are widely recorded, including in Los Angeles. Ten species were found in common with Los Angeles: Conicera similis, Dohrniphora cornuta, Megaselia albicaudata, M. basispinata, M. berndseni, M. nigra, M. pleuralis, M. ruficornis, M. rufipes and Spiniphora begenstammi. A similar faunal list from Henry Disney’s decades of multi-method trapping in Cambridge, UK (mostly in his backyard) can be found in Table 6 (R.H.L. Disney, Personal Communi-cation, May 24, 2016). This list shares nine of the same species with Los Angeles as the garden at Buckingham Palace (the Cambridge list also includes M. halterata but omits M. pleuralis). Interestingly, Disney’s observations on those species that will, at least on occasion, dominate sampling efforts did not include any of the overlap species we see in large numbers in Los Angeles (such as M. nigra). The dominating species in Cambridge did include two species (M. bendseni and M. rufipes) that we see (but in relatively low numbers) in Los Angeles.