Title Orchid bee responses to human disturbance in a tropxical forest are better detected using abundance and community redundancy. Data file: Allen_orchidbees_MLC_disturbance_data_2018.xlxs Creator Laura Allen, University of Glasgow Dates Data collected between June 2015 and September 2016. Data processed and analysed in 2017. Subject Orchid bee biodiversity along a human disturbance gradient in a tropical forest. Funders University of Glasgow, The Meyer Family Foundation, the Crees Foundation and the Darwin Initiative. Language English Location Alto Madre de Dios, Manu, Amazon, Peru, South America Methodology Orchid bees were sampled between 09:00-12:00 and 12:30-15:30 (+/- 15 mins for moving between sites). Each site was sampled for two morning and two afternoon sessions (a total of 12 hours per site). To reduce the effects of weather, two sites were sampled simultaneously - one more disturbed habitat (rank 1-3) and one less (rank 4-6), and different habitat types were sampled in the mornings and afternoons. No sampling was done if it was raining. At each site, two sampling stations were set up 50 m apart to reduce bias in the within-site location choice. At each of these stations, eight balls of cotton were hung from branches at a height of 1.5 m, with 2 m between cotton balls. Each cotton was baited with two drops of one of the following attractants: wintergreen oil, methyl salicylate, eucalyptus oil, eucalyptol, vanillin (3 tsp vanillin dissolved in 50 ml 96% ethanol), benzyl acetate, clove oil, eugenol. The bait stations were monitored over the three-hour sampling period, and orchid bees attracted to the baits were caught using hand nets and preserved in 70% ethanol. Orchid bees attracted to the survey area that did not settle at a specific bait were also captured. Orchid bees were later identified using published literature and museum reference specimens. An experiment on pollinator visitation was also conducted, with artificial flowers baited with sugar syrup used to attract pollinators to each site, and monitored over a total of 4.5 hours at each site. The number of visits made to the aritificial flowers were counted, included a note of any visits by orchid bees. Study site names: Sites: The site names contain two or three parts, i.e. 1.BA-A -> (1.)(BA)(-A). First, an arabic numeral that represents the disturbance level of the site (1-6), with 1 being the most disturbed category. For some purposes this number is ommitted in the site name and instead included as a separate variable, disturbance rank. Then the in the second part the letters describe the habitat category, as follows: BA - banana monoculture - banana plantations with minimal shade and AF - agroforestry plantation consiting of banana plants interspersed with native tree species used to provide shade for the crops as well as timber. SF - disturbed secondary forest lying in between small plantations, with dense understory and subject to regular ongoing disturbance. CCR - cleared regenerating forest, all cleared for cattle and crops, but regenerating under protection for >30 years. MXD - mixed history regenerating forest, with small patches cleared for crops and logging, regenerating under protectionfor >30 years. MIN - minimally disturbed tropical forest, with little evidence of logging and no known clearance, protected for >30 years. The least disturbed sites were contiguous with the surrounding forest, whereas the agricultural sites were limited in size by the boundary of the cultivated area. The smallest site, one of the banana plantations, was approximately 0.7 ha, but the rest were all over 1 ha. The third part, after the hyphen, indicates which replicate a site belongs to, with three sites sampled at each disturbance level. -A,-B or -C. All 'A' sites were surveyed during the same time period, and so on. Datasheets There are six data sheets included in this dataset. Each of these is described below: 1_OB_rawdata: This contains the list of all individual bees captures, the specimen numbers they have been given in the museum collections (specimens deposited Department of Entomology at the Universidad de San Antonio Abad de Cusco, and at the Museo de Historia Natural, Universidad Nacional Mayor de San Marcos, Lima), the site they were collected from (site codes explained above), their identification to genus and species level, whether they were captured in the morning or afternoon (am_pm), what bait they were attracted to, and the coordinates and elevation of the site they were collected from (in UTM and decimal degrees). 2_OB_count_matrix:The first column contains the species names of the orchid bees collected, and the first row lists the 18 sites from which samples were collected. 3_OB_diversity_variables: The data on the diversity of the orchid bee community caluclated for each site, combined with the data on the site location, the vegetation structure and the number of visits of pollinators to artificial flower baits. The variables are: Site Study site name River_distance Distance to the main river in m Road_distance Distance to the road in m Elevation Elevation above sea level in m UTM_Zone UTM_Coords_E UTM coordinates East UTM_Coords_S UTM coordinates South DD_Easting coordinates in decimal degrees DD_Northing Habitat Habitat type classification Rank Disturbance rank (1= most disturbed, 6= least) Site_rank Combined site-name rank column used for joining tables. veg_pc1 Output of vegetation structure PCA, component 1 veg_pc2 Output of vegetation structure PCA, component 2 Abundance_raw Raw count of orchid bee individuals collected Obs_alpha_q0 Observed alpha diversity at q = 0 Obs_alpha_q1 Observed alpha diversity at q = 1 Obs_alpha_q2 Observed alpha diversity at q = 2 Obs_alpha_qInf Observed alpha diversity at q = Infinity Est_alpha_q0 Estimated alpha diversity at q = 0 for n = 50 Est_alpha_q1 Estimated alpha diversity at q = 1 for n = 50 Est_alpha_q2 Estimated alpha diversity at q = 2 for n = 50 Redundancy_q0 Redundancy (rho) calculated at q = 0 Redundancy_q1 Redundancy (rho) calculated at q = 1 Redundancy_q2 Redundancy (rho) calculated at q = 2 Redundancy_q_Inf Redundancy (rho) calculated at q = Infinity Representativeness_q0 Representativeness (rho bar) calculated at q = 0 Representativeness_q1 Representativeness (rho bar)) calculated at q = 1 Representativeness_q2 Representativeness (rho bar) calculated at q = 2 Reresentativeness_qInfinity Representativeness (rho bar) calculated at q = Infinity Pollination_exp_visits Number of visits to an artificial flower array by potential pollinators Pollination_exp_orchidbee_visits Number of visits to an artificial flower array by orchid bees 4_vegetation_data_means: the mean of the vegetation measures collected at each site, used in the principal component analysis to assess how vegetation structure changes along the disturbance gradient. Details the same as for the raw vegation data, described below. 5_vegetation_data_raw: the raw values of the vegetation measurements collected at each study site. At each site, data were collected on the vegetation structure to assess how this changed across the disturbance gradient. Three vegetation points were randomly selected within each plot, and a circle of 10m diameter marked out at each. Within this circle, estimates of canopy height, canopy cover and understory vegetation density were collected, and tree counts and diameters and leaf litter depth measured. A single trained person should be assigned to estimate the height, and this person must do all points included in that survey to avoid bias. And average estimate of upper and mid canopy height should be recorded for each plot. At each of the corner and the centre of the plot, upper canopy cover (%) and mid canopy cover (%) were estimated using a frame. In the four outer points of the plot the following were recorded: a. The shrub layer: estimated looking outside of the survey point (NB – shrub layer lies between 1m and the base of the mid-canopy) using the modified Braun-Blanquet scale, defined below. b. The herb layer: estimated looking outside of the survey point (NB – from the ground to 1m) using the modified Braun-Blanquet scale below. In each quarter of the survey point, four points (16 points in total) will be randomly chosen. A ruler can be thrown within each quarter to choose these points, and a ruler/1m measuring stick used to take the following measurements: a. The leaf litter depth – the highest point at which the leaf touches the stick (nearest mm). b. If any herbs touch the stick. (a herb is classed as a plant <1m) c. If the stick hits bare ground. d. If the stick hits coarse woody debris. Within the circle we also recorded: a. The Diameter at Breast Height (DBH) of the three largest trees. b. The number of trees >5cm DBH. Modified Braun-Blanquet cover abundance scale used for herb and shrub layer densities: (as described by Allen & McLennan 1983; Allen 1993; Hurst &Allen 2007) 1 = <1% cover 2 = 1-5% cover 3 = 6-25% cover 4 = 26-50% cover 5 = 51-75% cover 6 = 76-100% cover 6_Weather_data_2009-2016 Weather data were collected daily at 0700, 1300 and 1900 daily. The current (actual) temperature and humidity were recorded at each of these times, as well as the maximum and minimum values since the last check. Rainfall was recorded at each time, and the total rainfall for the day also calculated. The weather recorder was reset after every recording, and when a record was missed (NA values), the time-sensitive data were omitted at the next session and then reset.