Satellite-based precipitation products (SPPs), such as the NOAA Climate Prediction Center (CPC) Morphing technique (CMORPH), have greatly extended our ability to monitor global precipitation. However, their performance over complex terrain remains highly uncertain. To improve accuracy, CPC has recently upgraded its operational SPP to the second generation CMORPH–CMORPH2. In addition to such efforts, the reliability of SPPs can be further enhanced by developing robust uncertainty estimates. This study employs a censored, shifted gamma distribution (CSGD)-based error modeling framework to develop error models for both CMORPH2 and its predecessor, CMORPH1, at a 6-hourly scale over Northern California—a typical complex terrain region that challenges most SPPs. Using the Stage IV reference precipitation, the relative improvements of CMORPH2 over CMORPH1 are quantified by comparing their CSGD-derived error statistics. The comparison shows that CMORPH2 outperforms CMORPH1 by reducing the overall bias and detection errors, and by better capturing the orographic gradients of precipitation over the Coast Ranges and Sierra Nevada. Validation results show that the trained error models can appropriately represent the bias and random errors; however, the models for both CMORPH2 and CMORPH1 show reduced skills in the high-altitude Sierra Nevada. With high-resolution regional climate simulations, the uncertainty estimates are further improved by incorporating them as covariates. The simulated precipitation shows substantial improvement of the uncertainty estimates over most areas, including the challenging high-altitude Sierra Nevada. Following precipitation, the simulated integrated water vapor transport (IVT) and convective available potential energy (CAPE) also offer modest improvements, mainly along the coast. Independent verification further demonstrates the robustness of the uncertainty estimates during heavy precipitation events. As NOAA is currently reprocessing CMORPH2 to produce an updated global precipitation record from 1991 onward, this error modeling framework holds potential for broader applications in quantifying the uncertainty of CMORPH2 for estimating orographic precipitation over extended periods and locations.

中文翻译：

---

表征 CMORPH 产品用于估计北加利福尼亚地形降水的不确定性


基于卫星的降水产品 （SPP），例如 NOAA 气候预测中心 （CPC） 变形技术 （CMORPH），极大地扩展了我们监测全球降水的能力。然而，它们在复杂地形上的表现仍然高度不确定。为了提高准确性，CPC 最近将其运营 SPP 升级到第二代 CMORPH-CMORPH2。除了这些努力之外，还可以通过开发稳健的不确定性估计来进一步提高 SPP 的可靠性。本研究采用基于删失、移位伽马分布 （CSGD） 的误差建模框架，在北加利福尼亚（一个典型的复杂地形区域，对大多数 SPP 构成挑战）上以 6 小时为CMORPH2及其前身 CMORPH1 开发误差模型。使用 IV 阶段参考降水，通过比较 CSGD 衍生的误差统计数据来量化 CMORPH2 相对于 CMORPH1 的相对改善。比较表明，CMORPH2 通过减少总体偏差和检测误差，以及更好地捕获海岸山脉和内华达山脉降水的地形梯度，优于 CMORPH1。验证结果表明，训练后的误差模型能够较好地表示偏差和随机误差;然而，CMORPH2 和 CMORPH1 的模型都显示高海拔内华达山脉的技能降低。通过高分辨率区域气候模拟，通过将不确定性估计作为协变量来进一步改进它们。模拟降水显示，大多数地区的不确定性估计有了显著改善，包括具有挑战性的高海拔内华达山脉。 降水后，模拟的综合水汽输送 （IVT） 和对流可用势能 （CAPE） 也提供了适度的改进，主要是沿着沿海地区。独立验证进一步证明了强降水事件期间不确定性估计的稳健性。由于 NOAA 目前正在对CMORPH2进行再处理，以生成 1991 年以来更新的全球降水记录，因此该误差建模框架在量化CMORPH2的不确定性以估计较长时间和位置的地形降水方面具有更广泛的应用潜力。