U.S. Climate Reference Network Research Articles

Correcting an Off-Nadir to a Nadir Land Surface Temperature Using a Multitemporal Thermal Infrared Kernel-Driven Model during Daytime

Land surface temperature (LST) is a fundamental parameter in global climate, environmental, and geophysical studies. Remote sensing is an essential approach for obtaining large-scale and frequently updated LST data. However, due to the wide field of view of remote sensing sensors, the observed LST with diverse view geometries suffers from inconsistency caused by the thermal radiation directionality (TRD) effect, which results in LST products being incomparable, especially during daytime. To address this issue and correct current off-nadir LSTs to nadir LSTs, a semi-physical time-evolved kernel-driven model (TEKDM) is proposed, which depicts multitemporal TRD patterns during the daytime. In addition, we employ a Bayesian optimization method to calibrate seven unknown parameters in the TEKDM. Validation results using the U.S. Climate Reference Network (USCRN) sites show that the RMSE (MBE) for GOES-16 and MODIS off-nadir LST products is reduced from 3.29 K (−2.0 K) to 2.34 K (−0.02 K), with an RMSE reduction of 0.95 K (29%) and a significant reduction in systematic bias. Moreover, the proposed method successfully eliminates the angular and temporal dependence of the LST difference between the satellite off-nadir LST and in situ nadir LST. In summary, this study presents a feasible approach for estimating the high-accuracy nadir LST, which can enhance the applicability of LST products in various domains.

On the Importance of Regime-Specific Evaluations for Numerical Weather Prediction Models as Demonstrated Using the High-Resolution Rapid Refresh (HRRR) Model

Abstract The scientific literature has many studies evaluating numerical weather prediction (NWP) models. However, many of those studies averaged across a myriad of different atmospheric conditions and surface forcings that can obfuscate the atmospheric conditions when NWP models perform well versus when they perform inadequately. To help isolate these different weather conditions, we used observations from the U.S. Climate Reference Network (USCRN) obtained between 1 January and 31 December 2021 to distinguish among different near-surface atmospheric conditions [i.e., different near-surface heating rates (), incoming shortwave radiation (SWd) regimes, and 5-cm soil moisture (SM05)] to evaluate the High-Resolution Rapid Refresh (HRRR) Model, which is a 3-km model used for operational weather forecasting in the United States. On days with small (large) , we found afternoon T biases of about 2°C (−1°C) and afternoon SWd biases of up to 170 W m−2 (100 W m−2), but negligible impacts on SM05 biases. On days with small (large) SWd, we found daytime temperature biases of about 3°C (−2.5°C) and daytime SWd biases of up to 190 W m−2 (80 W m−2). Whereas different SM05 had little impact on T and SWd biases, dry (wet) conditions had positive (negative) SM05 biases. We argue that the proper evaluation of weather forecasting models requires careful consideration of different near-surface atmospheric conditions and is critical to better identify model deficiencies in order to support improvements to the parameterization schemes used therein. A similar, regime-specific verification approach may also be used to help evaluate other geophysical models. Significance Statement Improving weather forecasting models requires careful evaluations against high-quality observations. We used observations from the U.S. Climate Reference Network (USCRN) and found that the performance of the High-Resolution Rapid Refresh (HRRR) Model varies as a function of differences in near-surface heating and solar radiation. This finding indicates that model evaluations need to be conducted under varying near-surface weather conditions rather than averaging across multiple weather types. This new approach will allow for model developers to better identify model deficiencies and is a useful step to helping improve weather forecasts.

Urbanization Effects in Estimating Surface Air Temperature Trends in the Contiguous United States

In the past century, local-scale warming caused by a strengthening urban heat island effect has brought inevitable systematic bias to observational data from surface weather stations located in or near urban areas. In this study, the land use situation around U.S. Climate Reference Network (USCRN) stations was used as a reference for rural station selection; stations with similar environmental conditions in the U.S. Historical Climatology Network (USHCN) were selected as reference stations using a machine learning method, and then the maximum surface air temperature (Tmax) series, minimum surface air temperature (Tmin) series and mean surface air temperature (Tmean) series of rural stations during 1921–2020 were compared with those for all nearby stations (including both rural and urban stations) to evaluate urbanization effects in the USHCN observation data series of the contiguous United States, which can be regarded as urbanization bias contained in the latest homogenized USHCN observation data. The results showed that the urbanization effect on the Tmean trend of USHCN stations is 0.002 °C dec−1, and the urbanization contribution is 35%, indicating that urbanization around USHCN stations has led to at least one-third of the overall warming recorded at USHCN stations over the last one hundred years. The urbanization effects on Tmax and Tmin trends of USHCN stations are −0.015 °C dec−1 and 0.013 °C dec−1, respectively, and the urbanization contribution for Tmin is 34%. These results have significance for understanding the systematic bias in USHCN temperature data, and they provide a reference for subsequent studies on data correction and climate change monitoring.

Diurnal to Decadal Variability in Land Surface and Air Temperature Gradient from 2002 to 2022 over the Contiguous United States

Abstract The surface and air temperature gradient (TS00 − Tair) drives the development of the convective boundary layer and the occurrence of clouds and precipitation. However, its variability is still poorly understood due to the lack of high-quality observations. This study fills in this gap by investigating the diurnal to decadal variability in TS00 − Tair from 2002 to 2022 based on hourly observations collected at over 100 stations of the U.S. Climate Reference Network. It is found that TS00 − Tair reaches its maximum at noon with an average of 6.85°C over the contiguous United States, which decreases to 4.28°C when the soil moisture exceeds 30%. The daily minimum of TS00 − Tair has an average of −2.08°C, which generally occurs in the early evening but is postponed as the cloud fraction decreases. Moreover, while existing studies have used the near-surface soil temperature, such as the 5-cm soil temperature (TS05), to calculate TS05 − Tair, we find that TS00 − Tair and TS05 − Tair have opposite diurnal cycles, and their amplitudes differed drastically. The daily minimum of TS00 − Tair has a significant decreasing trend (−0.50° ± 0.007°C decade−1) from 2002 to 2022 due to Tair increasing at a higher rate than TS00 during the nighttime. The occurrence frequency of near-surface stable conditions (TS00 − Tair < 0) increases significantly, and the frequency of unstable conditions (TS00 − Tair > 0) decreases notably throughout the year except for winter. When it is stable, the magnitude of TS00 − Tair tends to decrease while the TS00 − Tair tends to increase when it is unstable, which is consistent with the drying condition caused by the precipitation deficit. This study provides the first observational evidence on how TS00 − Tair responds to warming. Significance Statement The impact of global warming on surface-air temperature gradients is a crucial scientific issue that needs to be addressed. These gradients determine changes in cloud and precipitation, affecting water resources. However, traditional surface temperature measurements from weather stations are of high uncertainty due to direct exposure to the insolation. Satellite retrieval of surface temperature is limited by the availability of clear sky conditions, with low accuracy for temperature gradient calculations. Despite their importance, high-accuracy data of land surface temperature are still lacking. To address this issue, the U.S. Climate Reference Network (USCRN) uses infrared radiometers to continuously monitor surface temperature with high accuracy and sampling frequency. This study reports on surface-air temperature gradients at more than 100 U.S. stations, providing insight into diurnal to decadal variability over the contiguous United States. The study also highlights the significant difference between the land surface temperature–air temperature gradient and soil temperature–air temperature gradient.

Assessment of NA‐CORDEX regional climate models, reanalysis and in situ gridded‐observational data sets against the U.S. Climate Reference Network

AbstractClimate models' capability of reproducing the present climate at both global and regional scales still needs improvements. The assessment of model performance critically depends on the data sets used as comparators/references. Reanalysis and gridded observational data sets have been frequently used for this purpose. However, none of these can be considered an accurate reference data set because of their associated uncertainties and full representativity. This paper, for the first time, uses in‐situ measurements from National Oceanic and Atmospheric Administration U.S. Climate Reference Network (USCRN) spanning the period 2006–2020 to assess daily temperature and precipitation from a suite of dynamically downscaled regional climate models (RCMs; driven by ERA‐Interim) involved in NA‐CORDEX. The assessment is also extended to the most recent and widely used Earth system reanalyses (ERA5, ERA‐Interim, MERRA2 and NARR) and a few in situ‐based gridded data sets (Daymet, PRISM, Livneh and CPC). Results show that biases for the different data sets are seasonally and subregionally dependent. On average, reanalysis and in situ‐based gridded data sets are warmer (with biases exceeding 0.3°C) than USCRN year‐round, while RCMs are colder (warmer) in winter (summer) with biases ranging from −0.5 (0.9)°C for RCMs at 0.44° to −0.2 (1.4)°C for CRCM5‐UQAM‐11. In situ‐based gridded data sets provide the best performance in most of the Contiguous United States (CONUS) regions compared to reanalyses and RCMs, but still have biases in regions such as the Western mountains and the Pacific Northwest. Furthermore, in most US subregions, reanalysis data sets do not outperform reanalysis‐driven RCMs. Likewise, for both reanalysis data sets and RCMs, temperature and precipitation biases vary considerably depending on the local orography, with larger temperature biases for coarser model resolutions and precipitation biases for reanalysis.

A new reference-quality precipitation gauge wind shield

Abstract. Gauge-based precipitation measurements suffer from undercatch due to the effects of wind, with solid-precipitation measurements especially susceptible to such errors. When it is snowing and windy, unshielded precipitation gauges can catch less than half of the amount of precipitation of a gauge that is protected from the wind. For this reason, the US Climate Reference Network (USCRN) developed a large, double-layer, wooden wind shield called the Small Double Fence Intercomparison Reference (SDFIR). In past studies, the SDFIR has been demonstrated to be the most effective wind shield in use in any weather or climate network, reducing solid-precipitation undercatch to less than 10 % in wind speeds up to 8 m s−1. However, the wooden SDFIRs are subject to decay, they are difficult to replace and maintain, and they hinder access to maintaining the gauge. For these reasons, a new precipitation gauge wind shield called the Low Porosity Double Fence (LPDF) has been developed for use in the USCRN. Tested at three separate sites chosen for prevalent windy and snowy weather, the precipitation measurements recorded within the LPDF compared well to the SDFIR. After more than 2 years of measurements, the total precipitation recorded by the LPDF at each individual site differed by ±1.2 %, and the total LPDF accumulation from all sites was 0.03 % greater than the SDFIR accumulation. For the measurement of solid precipitation, the LPDF-shielded measurements were statistically indistinguishable from those in the SDFIR shield, and the time series of accumulation from precipitation gauges shielded by the SDFIR and the LPDF were almost identical. This new wind shield is much smaller and easier to install and maintain than any other reference-quality wind shield for the measurement of solid precipitation and may be of use within other meteorological, hydrological, and climate networks. It could also serve as a secondary reference precipitation measurement for precipitation intercomparisons held in remote locations where the construction of a full-sized Double Fence Intercomparison Reference (DFIR) shield is not feasible.

On the Detection of Remotely Sensed Soil Moisture Extremes

Abstract Remotely sensed soil moisture observations provide an opportunity to monitor hydrological conditions from droughts to floods. The European Space Agency’s (ESA) Climate Change Initiative has released both Combined and Passive datasets, which include multiple satellites’ measurements of soil moisture conditions since the 1980s. In this study, both volumetric soil moisture and soil moisture standardized anomalies from the U.S. Climate Reference Network (USCRN) were compared with ESA’s Combined and Passive datasets. Results from this study indicate the importance of using standardized anomalies over volumetric soil moisture conditions as satellite datasets were unable to capture the frequency of conditions observed at the extreme ends of the volumetric distribution. Overall, the Combined dataset had slightly lower measures of soil moisture anomaly errors for all regions; although these differences were not statistically significant. Both satellite datasets were able to detect the evolution from worsening to amelioration of the 2012 drought across the central United States and 2019 flood over the upper Missouri River basin. While the ESA datasets were not able to detect the magnitude of the extremes, the ESA standardized datasets were able to detect the interannual variability of extreme wet and dry day counts for most climate regions. These results suggest that remotely sensed standardized soil moisture can be included in hydrological monitoring systems and combined with in situ measures to detect the magnitude of extreme conditions. Significance Statement This study examines how well soil moisture extremes, wet or dry, can be detected from space using one of the lengthiest remotely sensed soil moisture datasets. Comparisons with high-quality station data from the U.S. Climate Reference Network revealed the satellite datasets could capture the frequency of extreme conditions important for climate monitoring, but often missed the absolute magnitudes of the extremes. Future research should focus on how to combine satellite and station data to improve the detection of extreme values important for monitoring.

Evaluation of Seasonal Differences among Three NOAA Climate Data Records of Precipitation

Abstract Three satellite precipitation datasets—CMORPH, PERSIANN-CDR, and GPCP—from the NOAA/Climate Data Record program were evaluated in their ability to capture seasonal differences in precipitation for the period 2007–18 over the conterminous United States. Data from the in situ U.S. Climate Reference Network (USCRN) provided reference precipitation measurements and collocated atmospheric conditions (temperature) at the daily scale. Satellite precipitation products’ (SPP) performance with respect to cold season precipitation was compared to warm season and full-year analysis for benchmarking purposes. Considering an ensemble of typical performance metrics including accuracy, false alarm ratio, probability of detection, probability of false detection, and the Kling–Gupta efficiency (KGE) that combines correlation, bias, and variability, we found that the three SPPs displayed better performances during the warm season than during the cold season. Among the three datasets, CMORPH displayed better performance—smaller bias, higher correlation, and a better KGE score—than the two other SPPs on an annual basis and during the warm season. During the cold season, CMORPH showed the worst performance at higher latitudes over areas experiencing recurring snow or frozen and mixed precipitation. CMORPH’s performances were further degraded compared to PERSIANN-CDR and GPCP when considering freezing temperatures (T < 0°C) due to the inability to microwave sensors to retrieve precipitation over snow-covered surface. However, for cold rainfall events detected simultaneously by the satellite and the USCRN stations (i.e., conditional case), CMORPH performance noticeably improved but remained inferior to the two other datasets. The quantification of seasonal precipitation errors and biases for three satellite precipitation datasets presented in this work provides an objective basis for the improvement of rainfall retrieval algorithms of the next generation of satellite precipitation products.

Using the U.S. Climate Reference Network to Identify Biases in Near- and Subsurface Meteorological Fields in the High-Resolution Rapid Refresh (HRRR) Weather Prediction Model

Abstract The ability of high-resolution mesoscale models to simulate near-surface and subsurface meteorological processes is critical for representing land–atmosphere feedback processes. The High-Resolution Rapid Refresh (HRRR) model is a 3-km numerical weather prediction model that has been used operationally since 2014. In this study, we evaluated the HRRR over the contiguous United States from 1 January 2021 to 31 December 2021. We compared the 1-, 3-, 6-, 12-, 18-, 24-, 30-, and 48-h forecasts against observations of air and surface temperature, shortwave radiation, and soil temperature and moisture from the 114 stations of the U.S. Climate Reference Network (USCRN) and evaluated the HRRR’s performance for different geographic regions and land cover types. We found that the HRRR well simulated air and surface temperatures, but underestimated soil temperatures when temperatures were subfreezing. The HRRR had the largest overestimates in shortwave radiation under cloudy skies, and there was a positive relationship between the shortwave radiation mean bias error (MBE) and air temperature MBE that was stronger in summer than winter. Additionally, the HRRR underestimated soil moisture when the values exceeded about 0.2 m3 m−3, but overestimated soil moisture when measurements were below this value. Consequently, the HRRR exhibited a positive soil moisture MBE over the drier areas of the western United States and a negative MBE over the eastern United States. Although caution is needed when applying conclusions regarding HRRR’s biases to locations with subgrid-scale land cover variations, general knowledge of HRRR’s biases will help guide improvements to land surface models used in high-resolution weather forecasting models. Significance Statement Weather forecasters rely upon output from many different models. However, the models’ ability to represent processes happening near the land surface over short time scales is critical for producing accurate weather forecasts. In this study, we evaluated the High-Resolution Rapid Refresh (HRRR) model using observations from the U.S. Climate Reference Network, which currently includes 114 reference climate observing stations in the contiguous United States. These stations provide highly accurate measurements of air temperature, precipitation, soil temperature, and soil moisture. Our findings helped illustrate conditions when the HRRR performs well, but also conditions in which the HRRR can be improved, which we expect will motivate ongoing improvements to the HRRR and other weather forecasting models.

Uncertainties on Climate Extreme Indices Estimated From U.S. Climate Reference Network (USCRN) Near‐Surface Temperatures

AbstractChanges in the frequency of temperature extremes are often attributed to global warming. The recent availability of near‐surface temperature data records from reference networks, such as the U.S. Climate Reference Network (USCRN), enables the quantification of measurement uncertainties. Within an activity of the Copernicus Climate Change Service, the estimation of the measurement uncertainty has been provided for USCRN temperature data, using metadata made available by the National Oceanic and Atmospheric Administration (NOAA). In this paper, four climate extreme indices (Frost Days, Summer Days, Ice Days, Tropical Nights) and the related uncertainties are calculated for the period 2006–2020 from the USCRN data set and compared with traditional indices. Moreover, the asymmetric USCRN measurement uncertainties are propagated to estimate the uncertainties of climate indices. The comparison shows expanded uncertainties homogeneously distributed with the latitude and typically within 15 days per year for Frost Days and within 10 days for Ice Days, while smaller uncertainties are estimated for Summer Days and Tropical Nights, with values typically within six to seven days per year. Positive uncertainties are typically larger than negative ones for all the indices. The values of Frost and Ice Days with the related uncertainties for USCRN have also been compared with the corresponding values calculated from reanalyses data, showing differences typically within 60 days for median values, quite often smaller than USCRN and inconsistent within the related uncertainties, Overall, the results show that USCRN measurement uncertainties increase confidence in the estimation of climate extreme indices and decisions for adaptation.

Evaluation of GLDAS soil moisture seasonality in arid climates

ABSTRACT We evaluated the Global Land Data Assimilation System surface soil moisture product (GLDAS v. 2.1) against in situ soil moisture networks in arid climates in Australia and the United States, using common statistical metrics and seasonality metrics. Our results showed that GLDAS performed well (root mean square error (RMSE) = 0.100 m3/m3; unbiased RMSE (ubRMSE) = 0.060 m3/m3; correlation coefficient (R) = 0.555 on average) but systematically overestimated the soil moisture values (Bias = 0.067 m3/m3). The performance was better in Australian Oznet and the U.S. Climate Reference Network (USCRN), compared to the US Soil Climate Analysis Network (SCAN) network. In terms of seasonality, GLDAS soil moisture seasons were biased to start earlier; on average, drying and wetting transitions started 28 and 16 days earlier than in situ data, respectively. The end dates of GLDAS seasonal transitions showed good agreement with in situ data; the errors in transition timings were limited to within a week. This tendency is stronger in the US networks compared to the Australian network.

A field evaluation of the SoilVUE10 soil moisture sensor

AbstractThe U.S. Climate Reference Network (USCRN) has been engaged in ground‐based soil water and soil temperature observations since 2009. As a nationwide climate network, the network stations are distributed across vast complex terrains. Due to the expansive distribution of the network and the related variability in soil properties, obtaining site‐specific calibrations for sensors is a significant and costly endeavor. Presented here are three commercial‐grade electromagnetic sensors, with built‐in thermistors to measure both soil water and soil temperature, including the SoilVUE10 Time Domain Reflectometry (TDR) probe (hereafter called SP) (Campbell Scientific, Inc.), 50 MHz coaxial impedance dielectric sensor (model HydraProbe, Stevens Water Monitoring Systems, Inc.) (hereafter called HP), and the TDR‐315L Probe (model TDR‐315L, Acclima, Inc.) (hereafter called AP), which were evaluated in a relatively nonconductive loam soil in Oak Ridge, TN, from 2021 to 2022. The HP manufacturer‐supplied calibration equation for loam soils was used in this study. While volumetric water content data from HP and AP were 82–99% of respective gravimetric observations at 10 cm, data from SP were only 65–81% of respective gravimetric observations in the top 20‐cm soil horizon, where soil water showed relatively large spatial variability. The poor performance of the SP is likely due to poor contact between SP sensor electrodes and soil and the presence of soil voids caused by the installation method used, which itself may have caused soil disturbance.

Evaluation of the Bias in the Use of Clear-Sky Compared with All-Sky Observations of Monthly and Annual Daytime Land Surface Temperature

Abstract Satellite-derived observations of land surface temperature (LST) are being utilized in a growing number of land surface studies; however, these observations are generally obtained from optical sensors that exclude cloudy observations of the land surface. The impact of using only clear-sky observations of land surfaces on monthly and annual estimates of daytime LST over two U.S. Climate Reference Network (USCRN) sites was evaluated over five years with daily in situ LST observations available for all-sky (clear and cloudy) conditions. The in situ LST observations were obtained for the nominal daytime observations associated with the MODIS sensors on board the Terra and Aqua satellites and were identified as all-sky or clear-sky conditions by utilizing cloud information provided with the MODIS LST product. Both monthly/annual mean and monthly/annual maximum values of daytime LST were significantly different when only clear-sky values were utilized, in comparison with all-sky values. Monthly averaged differences between the mean clear- and all-sky daytime LST (dLST) values ranged from −0.1° ± 1.5°C for January to 5.6° ± 1.8°C for May. Annually averaged dLST values, over the five years of the study, were 2.58°C, and differences between the maximum values of clear- and all-sky daytime LST values were −1.03°C. Although significant differences between mean annual clear-sky and all-sky daytime LST values were more frequent than differences observed for the annual maximum daytime LST values, the results suggest that the exclusive use of either mean or maximum clear-sky daytime LST values is not advisable for applications in which the use of daytime all-sky LST values would be more applicable.

An observational study of precipitation types in the Alaskan Arctic

The effects of various precipitation types, such as snow, rain, sleet, hail and freezing rain, on regional hydrology, ecology, snow and ice surfaces differ significantly. Due to limited observations, however, few studies into precipitation types have been conducted in the Arctic. Based on the high-resolution precipitation records from an OTT Parsivel2 disdrometer in Utqiaġvik, Alaska, this study analysed variations in precipitation types in the Alaskan Arctic from 15 May to 16 October, 2019. Results show that rain and snow were the dominant precipitation types during the measurement period, accounting for 92% of the total precipitation. In addition, freezing rain, sleet, and hail were also observed (2, 4 and 11 times, respectively), accounting for the rest part of the total precipitation. The records from a neighbouring U.S. Climate Reference Network (USCRN) station equipped with T-200B rain gauges support the results of disdrometer. Further analysis revealed that Global Precipitation Measurement (GPM) satellite data could well characterise the observed precipitation changes in Utqiaġvik. Combined with satellite data and station observations, the spatiotemporal variations in precipitation were verified in various reanalysis datasets, and the results indicated that ECMWF Reanalysis v5 (ERA5) could better describe the observed precipitation time series in Utqiaġvik and the spatial distribution of data in the Alaskan Arctic. Modern-Era Retrospective analysis for Research and Applications, Version 2 (MERRA-2) overestimated the amount and frequency of precipitation. Japanese 55-year Reanalysis (JRA-55) could better simulate heavy precipitation events and the spatial distribution of the precipitation phase, but it overestimated summer snowfall. Citation : Yue H D, Dou T F, Li S T, et al. An observational study of precipitation types in the Alaskan Arctic. Adv Polar Sci, 2021, 32(4): 324-337, doi: 10.13679/j.advps.2021.0027

Multi-layer high-resolution soil moisture estimation using machine learning over the United States

The lack of proper understanding of multi-layer soil moisture (SM) profile (signals) remains a persistent challenge in sustainable agricultural water management and food security, especially during drought conditions. We develop a machine-learning algorithm using the concept of learning from patterns to estimate the multi-layer SM information in ungauged locations firmly based on local knowledge of the climatic and landscape controls. The Contiguous United States (CONUS) is clustered into homogeneous regions based on the association between SM and climate and landscape controls. Extreme Gradient Boosting (XGBoost) algorithm is applied to homogenous regions to capture the complex relationship between appropriate predictor variables and in-situ SM at multiple layers over the CONUS. Soil Moisture Active Passive (SMAP) Level 4 (L4) surface (0–5 cm) and rootzone (0–100 cm) SM along with climate and landscape datasets are used as predictor variables. In-situ multi-layer SM recorded by Soil Climate Analysis Network (SCAN), Snow Telemetry (SNOTEL), and U.S. Climate Reference Network (USCRN) networks are utilized as predictands. XGBoost models are then trained region-wise and layer-wise to estimate multi-layer SM information at 5, 10, 20, 50, and 100 cm depths (five layers) at 1-km spatial resolution. Results indicate that the predictor variables have varying levels of influence on SM with changing soil depth, and meteorological variables have the least importance. Validation at 79 independent locations indicates the multi-layer SM estimates successfully capture temporal dynamics of SM, with most locations achieving ubRMSE less than 0.04 m3/m3. The high-resolution SM estimates offer spatial sub-grid heterogeneity compared to SMAP L4 SM.

Information-based uncertainty decomposition in dual-channel microwave remote sensing of soil moisture

Abstract. The National Aeronautics and Space Administration (NASA) Soil Moisture Active-Passive (SMAP) mission characterizes global spatiotemporal patterns in surface soil moisture using dual L-band microwave retrievals of horizontal (TBh) and vertical (TBv) polarized microwave brightness temperatures through a modeled mechanistic relationship between vegetation opacity, surface scattering albedo, and soil effective temperature (Teff). Although this model has been validated against in situ soil moisture, there is a lack of systematic characterization of where and why SMAP estimates deviate from the in situ observations. Here, we assess how the information content of in situ soil moisture observations from the US Climate Reference Network contrasts with (1) the information contained within raw SMAP observations (i.e., “informational random uncertainty”) derived from TBh, TBv, and Teff themselves and with (2) the information contained in SMAP's dual-channel algorithm (DCA) soil moisture estimates (i.e., “informational model uncertainty”) derived from the model's inherent structure and parameterizations. The results show that, on average, 80 % of the information in the in situ soil moisture is unexplained by SMAP DCA soil moisture estimates. Loss of information in the DCA modeling process contributes 35 % of the unexplained information, while the remainder is induced by a lack of additional explanatory power within TBh, TBv, and Teff. Overall, retrieval quality of SMAP DCA soil moisture, denoted as the Pearson correlation coefficient between SMAP DCA soil moisture and in situ soil moisture, is negatively correlated with the informational uncertainties, with slight differences across different land covers. The informational model uncertainty (Pearson correlation of −0.59) was found to be more influential than the informational random uncertainty (Pearson correlation of −0.34), suggesting that the poor performance of SMAP DCA at some locations is driven by model parameterization and/or structure and not underlying satellite measurements of TBh and TBv. A decomposition of mutual information between TBh, TBv, and DCA soil moisture shows that on average 58 % of information provided by TBh and TBv to DCA estimates is redundant. The amount of information redundantly and synergistically provided by TBh and TBv was found to be closely related (Pearson correlations of 0.79 and −0.82, respectively) to the retrieval quality of SMAP DCA. TBh and TBv tend to contribute large redundant information to DCA estimates under surfaces or conditions where DCA makes better retrievals. This study provides a baseline approach that can also be applied to evaluate other remote sensing models and understand informational loss as satellite retrievals are translated to end-user products.

Global evaluation of gridded satellite precipitation products from the NOAA Climate Data Record program

AbstractThree satellite gridded daily precipitation datasets: PERSIANN-CDR, GPCP, and CMORPH, that are part of the NOAA/Climate Data Record (CDR) program are evaluated in this work. The three satellite precipitation products (SPPs) are analyzed over their entire period of record, ranging from over 20-year to over 35-year. The products inter-comparisons are performed at various temporal (daily to annual) and for different spatial domains in order to provide a detailed assessment of each SPP strengths and weaknesses. This evaluation includes comparison with in-situ data sets from the Global Historical Climatology Network (GHCN-Daily) and the US Climate Reference Network (USCRN). While the three SPPs exhibited comparable annual average precipitation, significant differences were found with respect to the occurrence and the distribution of daily rainfall events, particularly in the low and high rainfall rate ranges. Using USCRN stations over CONUS, results indicated that CMORPH performed consistently better than GPCP and PERSIANN-CDR for the usual metrics used for SPP evaluation (bias, correlation, accuracy, probability of detection, and false alarm ratio among others). All SPPs were found to underestimate extreme rainfall (i.e. above the 90th percentile) from about -20% for CMORPH to -50% for PERSIANN-CDR. Those differences in performance indicate that the use of each SPP has to be considered with respect to the application envisioned; from the long-term qualitative analysis of hydro-climatological properties to the quantification of daily extreme events for example. In that regard, the three satellite precipitation CDRs constitute a unique portfolio that can be used for various long-term climatological and hydrological applications.

Developing and Validating Heat Exposure Products Using the U.S. Climate Reference Network

AbstractExtreme heat is one of the most pressing climate risks in the United States and is exacerbated by a warming climate and aging population. Much work in heat health has focused only on temperature-based metrics, which do not fully measure the physiological impact of heat stress on the human body. The U.S. Climate Reference Network (USCRN) consists of 139 sites across the United States and includes meteorological parameters that fully encompass human tolerance to heat, including relative humidity, wind, and solar radiation. Hourly and 5-min observations from USCRN are used to develop heat exposure products, including heat index (HI), apparent temperature (AT), and wet-bulb globe temperature (WBGT). Validation of this product is conducted with nearby airport and mesonet stations, with reanalysis data used to fill in data gaps. Using these derived heat products, two separate analyses are conducted. The first is based on standardized anomalies, which place current heat state in the context of a long-term climate record. In the second study, heat events are classified by time spent at various levels of severity of conditions. There is no consensus as to what defines a heat event, so a comparison of absolute thresholds (i.e., ≥30.0°, 35.0°, and 40.0°C) and relative thresholds (≥90th, 95th, and 98th percentile) will be examined. The efficacy of the product set will be studied using an extreme heat case study in the southeastern United States. While no heat exposure metric is deemed superior, each has their own advantages and caveats, especially in the context of public communication.

Using Ancillary Information from Radar-based Observations and Rain Gauges to Identify Error and Bias

AbstractAncillary information that exists within rain gauge and radar-based data sets provides opportunities to better identify error and bias between the two observing platforms as compared to error and bias statistics without ancillary information. These variables include precipitation type identification, air temperature, and radar quality. There are two NEXRAD based data sets used for reference; the National Centers for Environmental Prediction (NCEP) stage IV and the NOAA NEXRAD Reanalysis (NNR) gridded data sets. The NCEP stage IV data set is available at 4km hourly and includes radar-gauge bias adjusted precipitation estimates. The NNR data set is available at 1km at 5-minute and hourly time intervals and includes several different variables such as reflectivity, radar-only estimates, precipitation flag, radar quality indicator, and radar-gauge bias adjusted precipitation estimates. The NNR data product provides additional information to apply quality control such as identification of precipitation type, identification of storm type and Z-R relation. Other measures of quality control are a part of the NNR data product development. In addition, some of the variables are available at 5-minute scale. We compare the radar-based estimates with the rain gauge observations from the U.S. Climate Reference Network (USCRN). The USCRN network is available at the 5-minute scale and includes observations of air temperature, wind, and soil moisture among others. We present statistical comparisons of rain gauge observations with radar-based estimates by segmenting information based on precipitation type, air temperature, and radar quality indicator.

A Comparison of the U.S. Climate Reference Network Precipitation Data to the Parameter-Elevation Regressions on Independent Slopes Model (PRISM)

AbstractSince drought and excessive rainfall can have significant socioeconomic impacts, it is important to have accurate high-resolution gridded datasets that can help improve analysis and forecasting of these conditions. One such widely used dataset is the Parameter-Elevation Regressions on Independent Slopes Model (PRISM). PRISM uses a digital elevation model (DEM) to obtain gridded elevation analyses and then uses a regression analysis along with approximately 15 000 surface precipitation measurements to produce a 4-km resolution daily precipitation product over the conterminous United States. The U.S. Climate Reference Network (USCRN) consists of 114 stations that take highly accurate meteorological measurements across all regions of the United States. A comparison between the USCRN and PRISM was performed using data from 2006 to 2018. There were good comparisons between the two datasets across nearly all seasons and regions; most mean daily differences were <1 mm, with most absolute daily differences ~5 mm. The most general characteristics were for a net dry bias in the PRISM data in the Southwest and a net moist bias in the southern United States. Verifying the PRISM dataset provides us with confidence it can be used with estimates of evapotranspiration, high-resolution gridded soil properties, and vegetation datasets to produce a daily gridded soil moisture product for operational use in the analyses and prediction of drought and excessive soil moisture conditions.