**3. Results**

## *3.1. Survey Data*

The distribution of the number of colonies managed by beekeepers showed that most of the participating beekeepers (50.8%) had between 1–10 colonies (Figure 1A). On the contrary, only 1.7% of the beekeepers had more than 150 colonies but owned 21.5% of the colonies in the survey (Figure 1B). The distribution of losses per individual operation showed a positive skew with more than half (51.4%) of the participating beekeepers loosing between 0–10% of their colonies over the winter period and only 30.3% loosing more than 20% (Figure 1C). The approximate location of the main winter apiary (Figure 1D) showed a nationwide coverage all over Austria, with some areas more dominant than others, which could be traced back to geographical inaccessible areas like mountain ranges. Most participants did use the online form (*n* = 1378, 90%), while the paper form (*n* = 93, 6%) and the questionnaire from the beekeeping journal (*n* = 63, 4%) only accounted for 10% of the responses. The loss rate was not significantly different between online form users 15.4% (95% CI: 14.5–16.3%, *n* = 1378) and users who responded via paper 13.2% (95% CI: 10.4–16.5%).

**Figure 1.** (**A**,**B**) Operation size distribution in the survey: The operation size is the number of colonies owned by the individual beekeeper or beekeeping operation. (**C**) Distribution of losses in the survey, grouped into 10% loss rate groups. (**D**) Approximate location of the main winter apiary location: dot size represents number of apiaries in this area.

#### *3.2. Loss Rate Overview*

The total colony mortality over the winter period of 2018/2019 in Austria equals 15.2% (95% CI: 14.4–16.1%) (Figure 2A). The survey received valid answers from 1534 beekeepers, which stated that they wintered 33,651 colonies and had lost 5293 colonies over the winter (Table 1). Among these lost colonies, 1304 colonies were reported to struggle due to unsolvable queen-related problems and were counted as losses. Additionally, 60 colonies were lost due to natural disasters, which are not included in our risk analysis. The loss rate of the state Carinthia with 11.5% (95% CI: 9.5–14.0%) is significantly lower than the Austrian average. Burgenland had the lowest loss rate with 9.9% (95% CI: 6.0–15.7%) (Figure 2B) but also had the lowest number of participants (*n* = 35) and a wide confidence interval. Vienna had the highest loss rate but also the widest confidence interval with 19.6% (95% CI: 14.8–25.4%). The loss rates at the district level were highest for Waidhofen an der Ybbs, Lower Austria with 65.0% (95% CI: 48.6–79.9%, *n* = 7) and lowest for Tamsweg, Salzburg 0.0% (95% CI: 0.0–100.0%, *n* = 6).

The elevation of the main winter apiary location showed significantly lower loss rates for higher elevations (601–800 m: 13.0% (95% CI: 11.3–15.0%), >800 m: 11.8% (95% CI: 9.9–14.0%)) than the middle groups (201–400 m: 17.0% (95% CI: 15.4–18.9%), 401–600 m: 16.6% (95% CI: 15.0–18.2%)). No significant difference can be seen if compared to the lowest elevation group (0–200 m) with 15.4% (95% CI: 12.8–18.4%) (Figure 3).

**Figure 2.** (**A**) Austria and its state winter honey bee colony loss rates (and 95% CI) over the winter period of 2018/2019. (**B**) Map with loss rates colour coded on the Austrian state level. (**C**) Map with loss rates colour coded on the district level: white spots are *n* < 6.

**Figure 3.** Winter honey bee colony loss rates (and 95% CI) grouped into five different elevation above sea level categories.

## *3.3. Queen Management*

Colonies, which were still alive post winter but had queen problems that could not be solved by the beekeeper (drone laying queens or no queen at all) were counted as queen-related losses. This accounted for less than half of the combined losses, with an overall loss rate in Austria of 3.9% (95% CI: 3.6–4.1%). Lower Austria, 4.8% (95% CI: 4.3–5.4%), showed a significantly higher queen-related loss rate than the Austrian average (Figure 4A).

The distribution of the survey data showed that most beekeepers (85.1%) did renew more than a quarter of their queens (Figure 4B) and only 14.9% exchanged 0–25%. These result in a young queen percentage in the winter colonies population of about 52.3%. Loss rate without queen-related losses

in relation to the relative number of exchanged young queens, categorised in four groups, showed a significant higher loss rate for the first two groups with a lower number of renewed queens (0–25%: 17.1% (95% CI: 14.2–20.5%), 26–50%: 12.7% (95% CI: 11.4–14.3%)). The third group (51–75%) showed a significantly lower loss rate 10.1% (95% CI: 8.8–11.7%) compared to the first group, and the last group (75–100%) with most renewed queens had the lowest loss rate with 9.2% (95% CI: 7.4–11.3%) and was significantly lower than the first two groups (Figure 4C). Colony losses due to queen-related problems, using the same four categories of young queens, showed again a significantly higher loss rate for the first group with a 0–25% exchange rate and with 5.6% (95% CI: 4.7–6.6%) loss rate than the other three groups (26–50%: 3.8% (95% CI: 3.5–4.3%), 51–75%: 3.7% (95% CI: 3.3–4.1%), and 76–100%: 3.3% (95% CI: 2.7–3.9%)) (Figure 4D).

Participants were asked to what extent they observed queen problems in their colonies during the foraging season compared to previous seasons (four categories: more often, normal, more rare, and do not know). Only 7.8% of the participants stated that problems occurred more often, and 80.1% reported less or a normal experience (Figure 5A). There was no significant difference between these categories to loss rate without queen-related losses (Figure 5B). If looking at the queen-related losses in relation to the given categories, the category, "more often" had a significantly higher loss rate with 6.0% (95% CI: 4.8–7.3%) than "more rare" with 3.7% (95% CI: 3.1–4.3%) (Figure 5C).

**Figure 4.** (**A**) Queen-related losses (and 95% CI) for Austria and its states. (**B**) Distribution of survey participants who replaced their old queens in four categories. ( **C**) Categorized relative number of young queens and their winter honey bee colony loss rates (and 95% CI) without queen-related losses. (**D**) Categorized relative number of young queens and queen-related loss rate (and 95% CI).

**Figure 5.** (**A**) Rating of observed queen problems during the foraging season in four categories compared to previous season(s). (**B**) Winter honey bee colony loss rates (and 95% CI) without queen-related losses to queen problem occurrences. (**C**) Queen-related loss rate (and 95% CI) related to queen problem occurrences.

#### *3.4. Hive Management Practices*

Nine operational hive managemen<sup>t</sup> practices were provided with "yes", "no", or "uncertain" as possible options and compared as a single factor to colony mortality (Figure 6). Participants who did migrate their colonies had a significantly lower loss rate 11.3% (95% CI: 10.1–12.7%) than the ones who did not migrate their colonies 17.3% (95% CI: 16.3–18.5%) (Figure 6B). If beekeepers purchased wax from outside their own operation, they had a significantly higher loss rate 17.4% (95% CI: 15.8–19.2%) than participants with their own wax 14.0% (95% CI: 13.0–15.1%) (Figure 6G). No other significant effect of an operational factor on winter loss rate was found (Figure 6). The highest number of participants answering "uncertain" (*n* = 124, 9.3%) was recorded for the question whether small brood cell size was used.

Another question concerned the amount of old brood frames that were exchanged in the previous summer (Figure 7). The four categories in the survey were compared, which provided us with the relative amount of renewed brood frames by the participants. Higher exchange rates (>30%) showed a trend to lower loss rate but with no significant difference between the categories (Figure 7).

Reports categorized by operation size showed a significantly higher loss rate for beekeepers with 1–25 colonies (17.5%, 95% CI: 16.0–19.1%) over participants with 26–50 colonies (13.0%, 95% CI: 11.4–14.7%) and over 50 colonies (14.5%, 95% CI: 13.3–15.9%) (Figure 8).

**Figure 6.** Winter honey bee colony loss rates (and 95% CI) for nine different hive managemen<sup>t</sup> practices with three possible answers: "yes", "no", or "uncertain". These were compared as single factors to the loss rates. The category "uncertain" was removed from the plot when there were less than 30 answers.

**Figure 7.** Comparison between relative amount of exchange rate of old brood frames in four given percentage groups and corresponding winter honey bee colony loss rates (and 95% CI).

**Figure 8.** Categorized operation size groups and their winter honey bee colony loss rates (and 95% CI): The number of reported colonies before the onset of winter (2018) were used for the grouping.

#### *3.5. Forage as Risk Factor*

Participants were asked in the survey if their bees foraged on the following sources in 2018: oilseed rape (*Brassica napus*), maize (*Zea mays*), sunflower (*Helianthus annuus*), late catch crop, honeydew, and melezitose.

Significantly higher loss rates were observed with melezitose (yes: 17.4% (95% CI: 15.4–19.5%), no: 14.1% (95% CI: 13.1–15.3%)) (Figure A1F) and late catch crop (yes: 15.9% (95% CI: 14.5–17.3%, *χ*2 = 29.2, *p* < 0.05), no: 13.4% (95% CI: 12.0–15.0%)) (Figure A1D), and colonies with bees reported foraging in maize fields (yes: 18.1% (95% CI: 15.0–21.8%), no: 13.8% (95% CI: 12.7–14.9%)) (Figure A1B).

No difference was observed for rapeseed (yes: 16.5% (95% CI: 14.7–18.6%), no: 14.6% (95% CI: 13.5–15.8%)) (Figure A1A), sunflower (yes: 15.9% (95% CI: 13.9–18.1%), no: 14.9% (95% CI: 13.8–16.1%)) (Figure A1C), and honeydew (yes: 13.9% (95% CI: 12.5–15.5%), no: 15.5% (95% CI: 14.2–16.8%)) (Figure A1E).

The spatial distribution of the crops as reported by beekeepers revealed an aggregation of oil seed rape in Upper Austria and Lower Austria (Figure A1G). Most sunflower fields were present in Lower Austria (Figure A1I). Late catch crop and honeydew occurred all over Austria (Figure A1J,K). Melezitose occurred predominantly in Styria and Carinthia (Figure A1L).

#### *3.6. Varroa Control*
