Prevalent north–south striping (NSS) noise in the spherical harmonic coefficient products of the satellite missions gravity recovery and climate experiment greatly impedes the interpretation of signals. The overwhelming NSS noise always leads to excessive smoothing of the data, allowing a large room for improvement in the spatial resolution if this particular NSS noise can be mitigated beforehand. Here, we put forward a new spatial filter that can effectively remove NSS noise while remaining orthogonal to physical signals. This new approach overcomes the limitations of the previous method proposed by Swenson and Wahr (2006), where signal distortion was large and high-order coefficients were uncorrectable. The filter is based on autocorrelation in the longitude direction and cross-correlation in the latitude direction. The NSS-type noise identified by our method is mainly located in coefficients of spherical harmonic order larger than about 20 and degree beyond 30, spatially between latitudes ± 60°. After removing the dominating NSS noise with our method, a weaker filter than before is added to handle the residual noise. Thereby, the spatial resolution can be increased and the amplitude damping can be reduced. Our method can coincidentally reduce outliers in time series without significant trend bias, which underpins its effectiveness and reliability.

卫星任务重力恢复和气候实验的球谐系数产品中普遍存在的南北条纹（NSS）噪声极大地阻碍了信号的解释。压倒性的 NSS 噪声总是会导致数据过度平滑，如果可以事先减轻这种特定的 NSS 噪声，则空间分辨率有很大的改进空间。在这里，我们提出了一种新的空间滤波器，可以有效地去除 NSS 噪声，同时保持与物理信号的正交性。这种新方法克服了 Swenson 和 Wahr (2006) 提出的先前方法的局限性，其中信号失真很大，高阶系数无法校正。该滤波器基于经度方向的自相关和纬度方向的互相关。我们方法识别出的NSS型噪声主要位于球谐系数大于约20且度数超过30的区域，空间纬度在±60°之间。在用我们的方法去除了主要的 NSS 噪声后，添加了一个比以前更弱的滤波器来处理残余噪声。由此，可以提高空间分辨率，降低振幅阻尼。我们的方法可以巧合地减少时间序列中的异常值，而不会出现明显的趋势偏差，这是其有效性和可靠性的基础。可以提高空间分辨率，降低幅度阻尼。我们的方法可以巧合地减少时间序列中的异常值，而不会出现明显的趋势偏差，这是其有效性和可靠性的基础。可以提高空间分辨率，降低幅度阻尼。我们的方法可以巧合地减少时间序列中的异常值，而不会出现明显的趋势偏差，这是其有效性和可靠性的基础。