**4. Discussion**

Our study revealed that, despite significant warming in the past century, there were no significant changes in the drought patterns and drought impacts on the wild blueberry fields of Maine in the past 71 years. We also found that the water conditions (dry or wet, as indicated by the SPEI) in the growing season have significant impacts on wild blueberry vegetation vigor (as indicated by the NDVI and EVI) and production. The long-term water conditions (the long-term average SPEI) have substantial significant impacts on wild blueberry crop vegetation vigor (vegetation indices: NDVI and EVI) as well as their production (yield) in Maine, rather than the water conditions (SPEI) of the current growing season. The impact of the water conditions on the vegetation indices was more consistent and significant compared to the impact on yield. Interestingly, we also found that the water conditions of the early growing season (April–June) might decide the fate of crop vegetation vigor and production of wild blueberry later in the season (July–August). We further found that, in terms of vegetation status, water conditions had little impact on the irrigated field. The water conditions indicated by the SPEI had no impact on the yield of the irrigated field, suggesting irrigation effectively alleviated the impact of water deficits on the yield of wild blueberries. Based on our analyses, we also

**<LHOGOEVDFUH**

found that satellite-based vegetation indices (NDVI and EVI) cannot be used to predict wild blueberry crop production. However, several previous studies found significant correlations between high-resolution spectroradiometer-based vegetation indices and yield in different crops, i.e., maize, wheat, and soybean [32–34]. This could be because the yield of the wild blueberry crop is more affected by other factors such as pollination rather than the vegetation vigor. Moreover, further research could be carried out to test using drone-based high-resolution data to predict the yield of wild blueberries.

The absence of an increasing trend of the drought index SPEI in the wild blueberry fields could be associated with a lack of significant change in precipitation patterns during the growing season [6,15,35]. Although the atmospheric temperatures increased significantly in this region and in the wild blueberry fields in the past century [4–6,10,11], the warming pattern and subsequently increased evapotranspiration [6] have not resulted in a significant increase in drought impact. The studied fields are in a temperate climate region and they experience relatively low temperatures. The increase in evapotranspiration due to warming in this region has possibly not pushed the ecosystems here into the range of severe water deficits.

The water conditions of a relatively long period (SPEI of more than 11–12 months) showed significant and substantial impacts on vegetation vigor and the yield of wild blueberry crops. This could be because wild blueberries are a crop with large perennial underground stem systems called rhizomes, which can store sugar and nutrients [12–14,25], and their health and yield could mostly be determined by the sugar accumulation of previous years and not only that of the current growing season. Although the aboveground parts of the wild blueberries are pruned to the ground every two years, the belowground rhizomes and roots remain for a long time. As a result, the sugar stored underground could govern the effect of precipitation on the crop over the long term [36,37]. The wild blueberry crop requires only an inch of water per week [38] and is regarded as a droughtresistant crop [13,14]. This could be because of the large water and sugar storage in their underground tissues. The underground storage may weaken the effect of current year water conditions on crop health and yield.

The water conditions certainly affect the vegetation status and vigor of wild blueberries. The vegetation greenness or vigor of wild blueberries during both prune and crop years is affected by atmospheric temperature and precipitation during the growing period [6]. Also, precipitation directly affect the soil moisture availability to crops [39]. Furthermore, soil moisture availability affect the nitrogen uptake and accumulation in plants, which consequently determines leaf photosynthetic capacity [40], growth and yield of crops. However, a previous study on the wild blueberries in Eastern Canada found no correlations between the climate variables of that region and wild blueberry yield [41]. Further studies and analyses are needed to establish high-resolution relationships among climate variables, vegetation vigor, and yield.

Here, the vegetation indices (EVI and NDVI) are not good predictors of the yield of wild blueberries. This could be because, besides vegetation status, wild blueberry yield during the crop year is affected by many other important factors [42], such as pollination, insect pests, weeds, and pathogens. Though it has been found that vegetation indices are strongly correlated with yield in some crops [34], it might not be the case for wild blueberries. Vegetation indices are correlated with leaf chlorophyll content and photosynthesis capacity [43], and might be related to the number of developed flower buds [44]. However, there are a lot of other factors such as pollinator activity, weed coverage, and fruit set ratio, which are important in determining yield but can not be predicted by vegetation indices.

Water conditions (indicated by the SPEI) during the early growing season (April–June) have a larger impact on the vegetation status and yield of wild blueberry crops compared to that of the later growing season (July–September). This could be related to pollination. Precipitation intensity and frequencies, along with temperature and wind velocity, during the pollination period (April–May) in crop year would affect the bee pollination, which significantly affects the wild blueberry yield in July and August [42,45–47]. In addition,

the availability of resources such as soil moisture and nutrients [42], determined by the precipitation and temperature [38], during fruit set and maturation (May–June) right after the pollination period ends, decides the fate of the final fruit production (July–August) [45–48].

Irrigation decoupled the relationship between the climatic water condition (SPEI) and yield. The positive relationship between SPEI and yield found in the non-irrigated Baxter field was not found in the irrigated Airport field, despite both fields being in the same location with same management practices (except irrigation). The positive correlation between the SPEI and the yield of the non-irrigated Baxter field suggests the importance of water conditions in determining yield and the need for effective irrigation practices to alleviate the impact of drought. The fields of the major wild blueberry region (which are mostly non-irrigated) showed similar patterns to the non-irrigated Baxter field. Thus, it suggests that the introduction of effective irrigation management practices might be useful to enhance the production of wild blueberries by mitigating drought. Additionally, wild blueberries respond more positively to precipitation frequency rather than total precipitation volume over the growing season [49]. Irrigation also decoupled the relationship between the SPEI and vegetation indices (EVI and NDVI), suggesting the positive effect of irrigation in mitigating the drought effects on vegetation vigor for wild blueberries. Meanwhile, the quadratic relationships between the SPEI and vegetation indices, as well as the SPEI and yield, suggest that when the optimum precipitation or water supply is reached, further increases in the water supply may have a negative effect on crop vigor and yield. Similar results were also reported between the EVI and precipitation from the wild blueberry fields in Downeast, Maine [6]. Hence, no overall significant differences in the vegetation indices or yield were observed between the irrigated Airport field and the non-irrigated Baxter field, but in drought years (e.g., 2003), the yield and EVI of the irrigated field were higher than that of the non-irrigated field.
