Data from: Harvester ant seed removal in an invaded sagebrush ecosystem: implications for restoration

A better understanding of seed movement in plant community dynamics is needed, especially in light of disturbance-driven changes and investments into restoring degraded plant communities. A primary agent of change within the sagebrush-steppe is wildfire and invasion by non-native forbs and grasses, primarily cheatgrass (Bromus tectorum). Our objectives were to quantify seed removal and evaluate ecological factors influencing seed removal within degraded sagebrush-steppe by granivorous Owyhee harvester ants (Pogonomyrmex salinus Olsen). In 2014, we sampled 76 harvester ant nests across 11 plots spanning a gradient of cheatgrass invasion (40-91% cover) in southwestern Idaho, USA. We presented seeds from four plant species commonly used in post-fire restoration at 1.5 and 3.0 m from each nest to quantify seed removal. We evaluated seed selection for presented species, monthly removal, and whether biotic and abiotic factors (e.g., distance to nearest nest, temperature) influenced seed removal. Our top model indicated seed removal was positively correlated with nest height, an indicator of colony size. Distance to seeds and cheatgrass canopy cover reduced seed removal, likely due to increased search and handling time. Harvester ants were selective, removing Indian ricegrass (Achnatherum hymenoides) more than any other species presented. We suspect this was due to ease of seed handling and low weight variability. Nest density influenced monthly seed removal, as we estimated monthly removal of 1,890 seeds for 0.25 ha plots with 1 nest and 29,850 seeds for plots with 15 nests. Applying monthly seed removal to historical restoration treatments across the western U.S. showed harvester ants can greatly reduce seed availability at degraded sagebrush sites; for instance, fourwing saltbush (Atriplex canescens) seeds could be removed in <2 months. Collectively, these results shed light on seed removal by harvester ants and emphasize their potential influence on post-fire restoration within invaded sagebrush communities.

我们仍需更深入地理解植物群落动态中的种子移动规律，尤其是在干扰驱动的群落变化以及退化植物群落修复实践的背景下。蒿灌草原（sagebrush-steppe）生态系统的主要干扰驱动因子为野火与非本土草本及阔叶杂草的入侵，其中以旱雀麦（Bromus tectorum）的入侵最为显著。本研究旨在量化退化蒿灌草原中食种子的奥怀希收获蚁（Pogonomyrmex salinus Olsen）对种子的移除量，并解析影响种子移除的生态因子。2014年，我们在美国爱达荷州西南部的11个样地内开展采样，覆盖了旱雀麦入侵盖度40%~91%的梯度样带，共采集76个收获蚁蚁巢。我们选取4种常用于火灾后植被恢复的植物种子，在每个蚁巢1.5米和3.0米处分别放置供试种子，以量化种子移除量。本研究分析了收获蚁对供试物种的种子选择偏好、月度种子移除量，并探究了生物与非生物因子（如距最近蚁巢的距离、环境温度）是否会对种子移除产生影响。最优模型结果显示，种子移除量与蚁巢高度——蚁群规模的指示指标——呈显著正相关。供试种子与蚁巢的距离以及旱雀麦冠层盖度会降低种子移除量，这可能是因为上述因子延长了收获蚁的搜寻与处理时间。收获蚁具有明显的种子选择偏好，其对印度落芒草（Achnatherum hymenoides）的种子移除量显著高于其余供试物种。我们推测该偏好源于该种子易于处理且重量变异度较低。蚁巢密度会影响月度种子移除量：我们估算得出，当0.25公顷样地内仅含1个蚁巢时，月度种子移除量为1890粒；而当样地内蚁巢数量达15个时，月度种子移除量可达29850粒。将月度种子移除量应用于美国西部既往的修复处理场景后发现，收获蚁可大幅降低退化蒿灌生境中的种子可获得性；例如，四翼滨藜（Atriplex canescens）的种子可在2个月内被完全移除。综上，本研究结果阐明了收获蚁对种子的移除作用，并强调了其对入侵型蒿灌群落内火灾后植被修复的潜在影响。