Spatial and temporal dynamics of an aphid-ladybug metapopulation on Yucca glauca

Background/Question/Methods Here we ask how plant density and the temporal sequence of colonization affect predator prey cycling and population trajectory lags. We followed populations of aphids Aphis yuccae (wingless and winged morphs), ladybugs (Coccinella septempuncuata, Hyppodamia convergens ), ants, and sap beetles on inflorescences of 107 Yucca glauca plants for 30 days (the duration of the flowering season). The spatial location of all flowering yuccas within our 0.75-acre study area was determined with 15cm accuracy using high definition Trimble GPS. Our study site was located in a short-grass prairie in the foothills region of Central Colorado. We report on the spatial and temporal dynamics of aphids and aggregate (summed across species and life stage) ladybug populations. Results/Conclusions Autocorrelation function (ACF) analysis of the total aphid population in our study area revealed significant cycling with a fundamental period (T) of 14 days. ACF analysis of ladybug data showed an identical T. Together the aphid and ladybug population show correlated (r=0.69,n=27,p<0.0005) oscillations where ladybugs lag three days behind aphids. The aphid populations on individual plants exhibited a range of behaviors from non-cycling, to cycling with T from 8 to 40 days. The population oscillations on individual plants were characterized by crash-recolonization dynamics. Population dynamics of aphids on individual yuccas appear to be driven by density of yucca plants and time of first colonization: maximum population densities of aphids significantly decreased with increasing density of yucca plants (R2=0.17,p<0.0005) and decreased with increasing time of first colonization (R2=0.287,p<0.0005). Interestingly, time of first colonization is also weakly, but significantly related to yucca density; plants in low density are colonized significantly earlier than plants in high density (R2=0.07,p=0.008). This is surprising because Ripley's K analysis showed that distribution of winged morphs is spatially random. Time of first colonization also appears to drive joined aphid-ladybug dynamics. We used lagged correlation analysis between the ladybug and aphid population trajectories to determine if the populations are lagged and cycle at the lag together. Using this analysis we separated populations that are not correlated at any lag, populations that are lagged and do not cycle together, and populations that are correlated at a lag and cycle together. Using a bonferroni-corrected one-way ANOVA (F=5.772,p=0.004) we show that populations that are correlated and cycle are colonized significantly earlier than populations that are correlated but do not cycle which are colonized earlier (but not significantly) than populations that are unrelated to each other.