Objective monitoring methods that reliably identify grazing impact are required for long-term sustainable management in the arid rangelands. In Australia such methods must contend with highly unpredictable rainfall and large paddocks incorporating spatially complex land types with differing grazing preferences. Retrospective analysis of data collected following very dry and very wet periods removes vegetation dynamics generated by lesser rainfall events and should increase our ability to separate grazing effects from seasonal variability. Two remote-sensing methods were tested for their capacity to quantify trends over 30 years in vegetation-cover dynamics on a pastoral lease in central Australia with a history of heavy grazing. Following destocking by 2002, one section became a conservation reserve and another transitioned to a research station. During drought, the Dynamic Reference Cover Method (DRCM) showed that ground-cover deficit was less negative on both areas towards the end of the study. This improvement was attributable to increased ground cover after removal of grazing, followed by a reduced, but variable, grazing intensity on the research station and the spread of an introduced palatable perennial grass. Ground-cover response following rainfall was highest in 2011. Likewise, increased ground cover meant that the percentage Cover Production Loss (%CPL) index, calculated using grazing gradient methods (GGMs), was considerably less than a decade earlier following similar rainfall. Results from an associated recovery index (R) were inconclusive. Landscape heterogeneity potentially affected calculation of cover deficit using DRCM but, because heterogeneity was stable over time, reported change between dry years reliably indicated a trend owing to grazing. Interpreting trend from successive %CPL values in wet periods was complicated on the research station by altered waterpoint locations being superimposed on pre-existing degradation; however, the method should be effective in large paddocks with stable waterpoint locations. Despite their limitations, both methods can assist in objectively judging the long-term sustainability of grazing practices in contrasting seasonal conditions.

干旱牧场的长期可持续管理需要可靠地识别放牧影响的客观监测方法。在澳大利亚，这种方法必须应对高度不可预测的降雨量和大型围场，其中包括空间复杂的土地类型和不同的放牧偏好。对非常干旱和非常潮湿时期收集的数据进行回顾性分析，消除了降雨量减少所产生的植被动态，并且应该提高我们将放牧影响与季节变化分开的能力。对两种遥感方法进行了测试，以评估其对澳大利亚中部具有大量放牧历史的牧场租地 30 年来植被覆盖动态趋势进行量化的能力。2002 年库存减少后，其中一个区域成为自然保护区，另一个区域则转变为研究站。在干旱期间，动态参考覆盖方法（DRCM）显示，在研究结束时，这两个地区的地面覆盖赤字的负面影响有所减轻。这一改善归因于取消放牧后地表覆盖增加，随后研究站放牧强度减少但可变，以及引入的可口多年生草的传播。2011 年降雨后地被的响应最高。同样，地被覆盖的增加意味着使用放牧梯度方法 (GGM) 计算的覆盖物生产损失百分比 (%CPL) 指数比十年前类似降雨后的要少得多。相关恢复指数 (R) 的结果尚无定论。景观异质性可能影响使用 DRCM 计算覆盖赤字，但由于异质性随着时间的推移保持稳定，报告的干旱年份之间的变化可靠地表明了由于放牧而产生的趋势。在研究站上，由于水点位置的改变叠加在先前存在的退化上，解释潮湿时期连续 %CPL 值的趋势变得很复杂；然而，该方法对于具有稳定水点位置的大型围场应该是有效的。尽管存在局限性，但这两种方法都可以帮助客观地判断不同季节条件下放牧实践的长期可持续性。