An evaluation of Guided Regularized Random Forest for classification and regression tasks in remote sensing

New Earth observation missions and technologies are delivering large amounts of data. Processing this data requires developing and evaluating novel dimensionality reduction approaches to identify the most informative features for classification and regression tasks. Here we present an exhaustive evaluation of Guided Regularized Random Forest (GRRF), a feature selection method based on Random Forest. GRRF does not require fixing a priori the number of features to be selected or setting a threshold of the feature importance. Moreover, the use of regularization ensures that features selected by GRRF are non-redundant and representative. Our experiments based on various kinds of remote sensing images, show that GRRF selected features provides similar results to those obtained when using all the available features. However, the comparison between GRRF and standard random forest features shows substantial differences: in classification, the mean overall accuracy increases by almost 6% and, in regression, the decrease in RMSE almost reaches 2%. These results demonstrate the potential of GRRF for remote sensing image classification and regression. Especially in the context of increasingly large geodatabases that challenge the application of traditional methods.     

Characterization of sudden and sustained base flow jump hydrologic behaviour in the humid seasonal tropics of the Panama Canal Watershed

Base flows are important for tropical regions with pronounced dry seasons, which are facing increasing water demands. Base flow generation, however, is one of the most challenging hydrological processes to characterize in the tropics. In many years during the May–December wet season in the Panama Canal Watershed (PCW), base flows in rivers abruptly increase. This increase persists until the start of the December–April dry season. Understanding this unusual base flow jump (BFJ) behaviour is critical to improve water provisioning in the seasonal tropics, especially during droughts and extended dry seasons. This study developed an integrated approach combining piecewise regression on cumulative average base flow and sensitivity analysis to calculate the timing and magnitude of BFJ. Rainfall, forest cover, mean land surface slope, catchment area, and estimated subsurface storage were tested as predictors for the occurrence and magnitude of the BFJs in seven subcatchments of the PCW. Sensitivity analysis on correlated predictors allowed ranking of predictor contributions due to isolated and cross-correlation effects. Correlations between observed BFJs and BFJs predicted by watershed and rainfall-related predictors were 0.92 and 0.65 for BFJ timing and magnitude, respectively. Forest cover was the second most significant predictor after cumulative rainfall for jump magnitude, owing to larger subsurface storage and groundwater recharge in forests than pastures. Catchments in the mountainous eastern PCW always generated larger jumps due to their higher rainfall and greater forest cover than the western PCW catchments. The cross-correlations between predictors contributed to more than 50% of the jump variances. The results demonstrate the importance of rainfall gradient and catchment characteristics in affecting the sudden and sustained BFJs, which can help inform land management decisions intended to enhance water supplies in the tropics. This study underscores the need for more research to further understand the hydrological processes involved in the BFJ phenomenon, including better BFJ models and field characterizations, to help improve tropical ecosystem services under a changing environment.     

Intercomparison of measurements of bulk snow density and water equivalent of snow cover with snow core samplers: Instrumental bias and variability induced by observers

Manually collected snow data are often considered as ground truth for many applications such as climatological or hydrological studies. However, there are many sources of uncertainty that are not quantified in detail. For the determination of water equivalent of snow cover (SWE), different snow core samplers and scales are used, but they are all based on the same measurement principle. We conducted two field campaigns with 9 samplers commonly used in observational measurements and research in Europe and northern America to better quantify uncertainties when measuring depth, density and SWE with core samplers. During the first campaign, as a first approach to distinguish snow variability measured at the plot and at the point scale, repeated measurements were taken along two 20 m long snow pits. The results revealed a much higher variability of SWE at the plot scale (resulting from both natural variability and instrumental bias) compared to repeated measurements at the same spot (resulting mostly from error induced by observers or very small scale variability of snow depth). The exceptionally homogeneous snowpack found in the second campaign permitted to almost neglect the natural variability of the snowpack properties and focus on the separation between instrumental bias and error induced by observers. Reported uncertainties refer to a shallow, homogeneous tundra-taiga snowpack less than 1 m deep (loose, mostly recrystallised snow and no wind impact). Under such measurement conditions, the uncertainty in bulk snow density estimation is about 5% for an individual instrument and is close to 10% among different instruments. Results confirmed that instrumental bias exceeded both the natural variability and the error induced by observers, even in the case when observers were not familiar with a given snow core sampler.     

Cover crops and precipitation influence soluble reactive phosphorus losses via tile drain discharge in an agricultural watershed

Subsurface tile drainage speeds water removal from agricultural fields that are historically prone to flooding. While managed drainage systems improve crop yields, they can also contribute tothe eutrophication of downstream ecosystems, as tile-drained systems are conduits for nutrients to adjacent waterways. The changing climate of the Midwestern US has already altered precipitation regimes which will likely continue into the future, with unknown effects on tile drain water and nutrient loss to waterways. Adding vegetative cover (i.e., as winter cover crops) is one approach that can retain water and nutrients on fields to minimize export via tile drains. In the current study, we evaluate the effect of cover crops on tile drain discharge and soluble reactive phosphorus (SRP) loads using bi-monthly measurements from 43 unique tile outlets draining fields with or without cover crops in two watersheds in northern Indiana. Using four water years of data (n = 844 measurements), we examined the role of short-term antecedent precipitation conditions and variation in soil biogeochemistry in mediating the effect of cover crops on tile drain flow and SRP loads. We observed significant effects of cover crops on both tile drain discharge and SRP loads, but these results were season and watershed specific. Cover crop effects were identified only in spring, where their presence reduced tile drain discharge in both watersheds and SRP loads in one watershed. Varying effects on SRP loads between watersheds were attributed to different soil biogeochemical characteristics, where soils with lower bioavailable P and higher P sorption capacity were less likely to have a cover crop effect. Antecedent precipitation was important in spring, and cover crop differences were still evident during periods of wet and dry antecedent precipitation conditions. Overall, we show that cover crops have the potential to significantly decrease spring tile drain P export, and these effects are resilient to a wide range of precipitation conditions.     

NOAA IMS雪冰产品在青藏高原积雪监测中的适用性分析

在系统评估青藏高原积雪观测典型气象站历史定位坐标精度基础上,利用站点雪深资料对NOAA IMS 4 km和1 km分辨率雪冰产品在青藏高原的精度和适用性进行了验证和评估,定量分析了IMS 4 km到1 km空间分辨率提高和气象站历史定位与GPS定位坐标之间的差异对青藏高原IMS积雪监测精度的影响。结果表明：青藏高原个别气象站历史坐标与当前GPS接收机定位之间存在较大的差异,如安多气象站经度偏小0.6°,纬度偏大0.08°。IMS 4 km雪冰产品在青藏高原的总精度介于76.4%~83.2%,平均为80.1%,积雪分类精度介于35.8%~60.7%,平均为47.2%,平均误判率为17.1%,平均漏判率为45.5%,总体上呈现地面观测的积雪日数越多、平均雪深越大,其总体监测精度越低,而积雪分类精度越高的特点。IMS分辨率从4 km到1 km总体精度平均提高了2.9%,积雪分类精度平均提高了0.9%,主要是由于个别站点的精度提升较大引起的,对高原多数台站积雪监测精度的改进和提升很小。除个别台站外,目前气象站历史坐标和GPS定位坐标之间的差异,对IMS 4 km积雪监测精度验证结果没有影响。然而,今后随着卫星遥感技术的发展,更高时空分辨率的遥感积雪产品将用于积雪监测和研究,精确的地面观测站坐标信息是对这些遥感数据开展精度验证与实际应用的前提。     

积雪变化对中国森林凋落物分解影响研究进展

森林凋落物的分解对于维持生态系统物质循环和养分平衡具有重要意义,并受到不同积雪厚度下冻融格局的影响。冻融期（包括冻结过程期、完全冻结期、融化过程期）是冻土区凋落物分解的重要时期,该时期分解的凋落物量约占全年分解总量的一半。积雪减少通常会导致土壤温度降低、冻融循环次数增加,进而影响凋落物分解。通过综述近10年来积雪变化对我国森林凋落物分解影响的研究成果发现,积雪厚度减少在冻融期通常会抑制凋落物质量损失、碳元素释放和纤维素降解,生长季则起到促进作用,从全年来看多数表现为抑制作用。因此,冻融作用造成凋落物的物理破坏,对其分解的促进作用主要发生在后续生长季。积雪厚度减少在冻融期通常抑制氮元素释放,生长季和全年则无明显规律;磷元素和木质素目前研究还存在很大差异。最后,进一步阐述了积雪变化对凋落物分解影响研究存在的问题及未来研究发展方向。     

长白山苔原带土壤温度与肥力随海拔的变化特征

土壤温度与土壤肥力的分解释放、植被生长密切相关。利用2015年8月至2017年6月长白山西坡苔原带5 cm土壤温度并测试其土壤肥力,分析了土壤温度与肥力随海拔的变化特征及土壤温度对苔原带肥力的影响。结果表明：（1）长白山西坡苔原带土壤最热月为8月,最冷月为1、2月。长白山西坡苔原带土壤年均温随海拔的升高而下降,垂直变化率为-0.44℃·（100m）^-1。月均温垂直变化率则有所差别,5-9月垂直变化率为正,其余月份垂直变化率为负。（2）海拔是土壤温度空间分异的主要影响因素,冷季土壤温度随海拔升高而显著降低。随着海拔升高,越稀疏的植被和越薄的土层使得土壤热容量越小,暖季土壤温度随海拔升高而显著升高。（3）长白山西坡苔原带土壤肥力,尤其是与植物生长关系密切的速效养分随海拔升高表现出先升高再降低,在植物多样性和丰富度及草本植物盖度最高的2 250 m处达到土壤肥力最高水平。低海拔（2 050~2 250 m）的土壤肥力水平明显高于高海拔（2 350~2 550 m）的土壤肥力水平。西坡苔原带土壤肥力的空间分异状况受草本植物入侵影响较大。（4）长白山西坡苔原带土壤肥力水平随土壤温度升高而升高,温度是土壤有机质分解和矿物质养分转化的限制性因素。建议山地苔原带生态系统生产和生态管理中要重点考虑草本植被入侵给土壤肥力带来的影响。     

南襄盆地南阳凹陷地热成因研究

通过研究南阳凹陷区域地质构造,地温场特征和热储层特征,认为南阳凹陷地热田为典型的沉积盆地型地热田.南阳凹陷热源主要由地下水在正常梯度下经围岩增温加热和地下水沿深大断裂的上涌两部分组成,朱阳关——夏馆——大河断裂、方城——邓县断裂、新野断裂等活动性深大断裂为热量传导提供了良好的通道,古近系稳定的砂岩泥岩互层为深层地热良好...     

基于MODIS数据的北疆积雪黑碳和雪粒径反演及时空变化分析

积雪是地球上反射率较高的自然表面,对于中高纬度地区的水文和能量收支平衡发挥着重要作用。表层积雪中的黑碳和雪粒径变化可以显著影响积雪反照率,造成积雪对太阳辐射吸收的变化,进而对区域气候变化和水文循环产生反馈作用。利用遥感技术对季节性积雪表层黑碳和雪粒径进行定量评估,可以获取时空上连续系统的雪表黑碳浓度和雪粒径变化情况,这也是许多气候和水文模型的输入因子。以中国主要季节性积雪区北疆为研究区,基于MODIS（Moderate Resolution Imaging Spectroradiometer）数据的3（0.47 μm）、2（0.86 μm）和5（1.24 μm）波段,采用SGSP（Snow Grain Size and Pollution Amount）算法反演2000-2018年积雪期的雪表黑碳浓度和雪粒径,并结合地面观测数据对于反演结果进行了精度验证,综合分析北疆雪表黑碳浓度和雪粒径时空变化趋势。结果显示,SGSP算法能够同时反演雪表黑碳浓度和雪粒径,并且验证结果表明纯雪像元上反演结果具有较好的精度;2000-2018年北疆雪表年均黑碳浓度和年均雪粒径都随时间变化呈现微弱下降趋势;受地理位置和局部污染源的影响,北疆积雪黑碳浓度空间分布复杂,天山北坡经济带平均黑碳浓度最高,伊犁地区平均黑碳浓度最低,雪粒径的空间分布显示塔城地区平均雪粒径最大,伊犁地区最小。     

赣南白面石地区铀资源潜力分析

白面石地区有着得天独厚的铀成矿地质背景,区内铀矿化是在富铀的白面石花岗岩体基础上发展起来的。富铀基底花岗岩的风化,使铀活化迁移,形成高铀含量砂岩层,是该区成矿的基础;北西向展布的双峰式火山岩浆喷溢、形成覆盖在高铀含量砂岩层之上的火山岩层,起着热盖和封闭作用,使铀聚积成矿,是成矿的关键;后期断裂作用及次火山岩的贯入,带来了丰富的热能,使铀再次活化迁移,铀矿化叠加变富。