Moisture Data Research Articles

Regionalization of GRACE data in shorelines by ensemble of artificial intelligence methods

Groundwater (GW) plays a crucial role in coastal aquifers and arid regions, serving as a lifeline for communities by providing a reliable and resilient water source, making its monitoring essential for sustainable water management. This study aimed at modeling GW via regionalization of the Gravity Recovery and Climate Experiment (GRACE) data based on two methods. The first method directly regionalized the GRACE data for modeling GW via in situ measurements, including the lake level, precipitation, temperature, observed GW and Penman-Monteith-Leuning (PML) evapotranspiration data. The second method included two stages, in the first stage, the GRACE data were downscaled via the Famine Early Warning Systems Network Land Data Assimilation System (FLDAS) data which contains satellite based precipitation, temperature, soil moisture, and snow water equivalent data. In the second stage, the downscaled GRACE was bias corrected to provide regionalized data. Artificial intelligence models consist of shallow networks (Feed Forward Neural Network (FFNN), Adaptive neuro fuzzy (ANFIS), Support Vector Machine (SVR)), the ensemble of shallow networks and Long-Short Term Memory (LSTM) deep learning method were employed in the modeling process and the observed GW level data were targeted for the regionalization. The Link CluE clustering ensemble method was implemented to cluster the piezometers of the aquifer to separate different GW patterns in the area. The proposed methodology was examined over the Miandoab plain, one of the sub-basins of the Lake Urmia, located in Northwest Iran. The modeling results demonstrated that the first method could exhibit superior performance with the Nash-Sutcliffe Efficiency (NSE) of up to 17% higher than the second method. Thus, using in situ observed data for downscaling proved to be more accurate than relying on the data based on the satellite imagery. The results indicated that the ensemble of shallow networks could lead to more precise results than using the deep and shallow learning models, individually, where the NSE for the ensemble of shallow networks was up to 50% higher compared to the LSTM model.

Use of Soil Moisture as an Indicator of Climate Change in the SUPer System

Soil moisture can be an important indicator of climate change in humid and semi-arid areas. This indicator can more efficiently propose different public policies related to climate change than just using precipitation and temperature data. Given the above, the objective of this study is to evaluate changes in soil moisture in the state of Pernambuco during the period 1961–2021, using the System of Hydrological Response Units for Pernambuco. In this study, two river basins in the state of Pernambuco that represent the different climatic conditions of the state were chosen. The results show that in the coastal region there is a tendency towards more saturated soils, and in the semi-arid region there is a tendency towards drier soils. With these results, it is possible to conclude that public policy decisions for the economy, environment, and society must consider this vital water balance variable. Leveraging soil moisture and precipitation data makes it possible to differentiate between flood risks and landslide vulnerabilities, particularly in regions characterized by higher levels of rainfall. Monitoring soil water content in humid and semi-arid areas can significantly enhance early warning systems, thereby preventing loss of life and minimizing the socio-economic impacts of such natural events. As such, this study provides a holistic understanding of the relationship between climatic patterns, soil moisture dynamics, and the occurrence of droughts and floods, ultimately contributing to more effective disaster preparedness and response measures in Pernambuco and similar regions.

Development and preliminary validation of a land surface image assimilation system based on the Common Land Model

Abstract. Data assimilation is an essential approach to improve the predictions of land surface models. Due to the characteristics of single-column models, assimilation of land surface information has mostly focused on improving the assimilation of single-point variables. However, land surface variables affect short-term climate more through large-scale anomalous forcing, so it is indispensable to pay attention to the accuracy of the anomalous spatial structure of land surface variables. In this study, a land surface image assimilation system capable of optimizing the spatial structure of the background field is constructed by introducing the curvelet analysis method and taking the similarity of image structure as a weak constraint. The fifth-generation ECMWF Reanalysis – Land (ERA5-Land) soil moisture reanalysis data are used as ideal observation for the preliminary effectiveness validation of the image assimilation system. The results show that the new image assimilation system is able to absorb the spatial-structure information of the observed data well and has a remarkable ability to adjust the spatial structure of soil moisture in the land model. The spatial correlation coefficient between the model surface soil moisture and observation increased from 0.39 to about 0.67 after assimilation. By assimilating the surface soil moisture data and combining these with the model physical processes, the image assimilation system can also gradually improve the spatial structure of soil moisture content at a depth of 7–28 cm, with the spatial correlation coefficient between the model soil moisture and observation increased from 0.35 to about 0.57. The forecast results show that the positive assimilation effect could be maintained for more than 30 d. The results of this study adequately demonstrate the application potential of image assimilation system in short-term climate prediction.

IoT Based Smart Irrigation System using Artificial Intelligence

This proposed of the project introduces an innovative IoT-based AI-driven solution engineered to revolutionize water management in agricultural contexts, catering to both large-scale piped irrigation networks and micro-irrigation systems. Our proposed system leverages cutting-edge AI algorithms to forecast dynamic crop water requirements, seamlessly integrating real-time soil moisture data to optimize water utilization and amplify crop yields. By amalgamating AI-driven predictive analytics with instantaneous sensor feedback, our system facilitates proactive water resource management, curbing wastage and mitigating environmental repercussions. Moreover, the integration of Deep Learning-based pest detection augments the system's capabilities by safeguarding crops against potential threats, thus fostering sustainable agricultural practices. Additionally, our solution pioneers the incorporation of Deep Learning-based crop production for the detection of wild animals, thereby enabling timely alerts through buzzer sound notifications, thereby ensuring comprehensive crop protection. This amalgamation of IoT, AI, and Deep Learning technologies promises to significantly enhance agricultural sustainability, optimize resource utilization, and mitigate ecological impact, thus paving the way for a more resilient and productive agricultural ecosystem.

Evaluation of the fully coupled WRF and WRF-Hydro modelling system initiated with satellite-based soil moisture data

ABSTRACT This study comparatively evaluates the skill of the (un)coupled Weather Research and Forecasting (WRF) – WRF-Hydro system supported with satellite-based soil moisture initiation in long-term simulations of hydrometeorological variables under humid and semi-arid climate conditions. Prior to the coupling, the WRF-Hydro model is calibrated using streamflow data at different data scales (i.e. hourly and daily) to discuss the influences of the changing temporal resolution of the reference dataset on the optimum parameter values. According to the results, fully coupled models correct the location and magnitude of the occurrence of hourly extreme precipitation. Changing the initial soil moisture data source affects the terrestrial water cycle elements in the semi-arid region (i.e. Mediterranean (MED)) more than in the humid area (i.e. Eastern Black Sea (EBS)), especially in spring and summer. The initiation of the coupled model with satellite-based soil moisture improves the streamflow outputs after a reasonable spin-up period for the MED region.

HiMIC-Monthly: A 1 km high-resolution atmospheric moisture index collection over China, 2003–2020

Near-surface atmospheric moisture is a key environmental and hydro-climatic variable that has significant implications for the natural and human systems. However, high-resolution moisture data are severely lacking for fine-scale studies. Here, we develop the first 1 km high spatial resolution dataset of monthly moisture index collection in China (HiMIC-Monthly) over a long period of 2003~2020. HiMIC-Monthly is generated by the light gradient boosting machine algorithm (LightGBM) based on observations at 2,419 weather stations and multiple covariates, including land surface temperature, vapor pressure, land cover, impervious surface proportion, population density, and topography. This collection includes six commonly used moisture indices, enabling fine-scale assessment of moisture conditions from different perspectives. Results show that the HiMIC-Monthly dataset has a good performance, with R2 values for all six moisture indices exceeding 0.96 and root mean square error and mean absolute error values within a reasonable range. The dataset exhibits high consistency with in situ observations over various spatial and temporal regimes, demonstrating broad applicability and strong reliability.

Advancing Spatial Drought Forecasts by Integrating an Improved Outlier Robust Extreme Learning Machine with Gridded Data: A Case Study of the Lower Mainland Basin, British Columbia, Canada

Droughts have extensive consequences, affecting the natural environment, water quality, public health, and exacerbating economic losses. Precise drought forecasting is essential for promoting sustainable development and mitigating risks, especially given the frequent drought occurrences in recent decades. This study introduces the Improved Outlier Robust Extreme Learning Machine (IORELM) for forecasting drought using the Multivariate Standardized Drought Index (MSDI). For this purpose, four observation stations across British Columbia, Canada, were selected. Precipitation and soil moisture data with one up to six lags are utilized as inputs, resulting in 12 variables for the model. An exhaustive analysis of all potential input combinations is conducted using IORELM to identify the best one. The study outcomes emphasize the importance of incorporating precipitation and soil moisture data for accurate drought prediction. IORELM shows promising results in drought classification, and the best input combination was found for each station based on its results. While high Area Under Curve (AUC) values across stations, a Precision/Recall trade-off indicates variable prediction tendencies. Moreover, the F1-score is moderate, meaning the balance between Precision, Recall, and Classification Accuracy (CA) is notably high at specific stations. The results show that stations near the ocean, like Pitt Meadows, have higher predictability up to 10% in AUC and CA compared to inland stations, such as Langley, which exhibit lower values. These highlight geographic influence on model performance.

Modelling and optimization of urea super granule (USG) placement depth in paddy cultivation under check basin irrigation using HYDRUS-2D model

Urea, a nitrogenous fertilizer which plays an important role in enhancing yield, but has a drawback of low nitrogen use efficiency (NUE). Use of urea super granule (USG) and coated urea have advantages over prilled urea in terms of higher NUE and eventually improved yield. Depth of placement (DOP) of USG plays a major role in achieving higher NUE and yield. However, appropriate depth of placement needs detailed knowledge about water and nitrogen (N) movements to provide maximum nutrient to plants. Therefore, an experiment was conducted in the paddy field using USG coated with three different types of coating (USG with nano bentonite and neem oil (T1), USG with heated nano bentonite and neem oil (T2) and USG with sulfer and acacia gum solution (T3)) and placed at the centre of four hills of paddy as basal dose at ICAR-IARI, New Delhi. Spatial and temporal data of soil moisture and Nwere collected for the initial 32 days in paddy field after crops establishment. The HYDRUS-2D model was calibrated and validated using the observed data. Calibrated model was used for simulation of N movement in different soil types and for different depths of USG placement. Results revealed that permissible soil like sandy loam, the optimum DOP irrespective of coating was 5 cm. For sandy clay loam soil, optimum depth of placement was found to be 10 cm, 6 cm and 10 cm for T1, T2 and T3 type of coatings. Similarly, for clay loam soil, depth of USG placement for T1, T2 and T3 type coating was found to be 8 cm, 10 cm and 8 cm, respectively. Using the optimum DOP in sandy clay loam soil, results showed that both i.e., DOP and coating type has significant effect on number of tillers m−2 and yield (t ha−1). In 2020 and 2021, the highest number of tillers per square meter at 30 days after transplanting (DAT) was observed in the T3 treatment at DOP of 10 cm, with counts of 453.66 and 454.88, respectively. Following closely were T2 at DOP of 10 cm (447.33 and 119.55) and T1 at DOP of 6 cm (443.44 and 444.77). Similarly, the best yield in 2020 and 2021 were observed from T3 treatment at DOP of 10 cm (5.41 t/ha and 5.44 t/ha), followed by T2 at DOP of 10 cm (5.33 t/ha and 5.42 t/ha), and T1 at DOP of 6 cm (5.29 t/ha and 5.31 t/ha). This study may assist the farmers, researchers, manufacturers and policy makers to make selection of the optimum depth of placement for higher NUE and yield in different soil types.

Spatial Downscaling of ESA CCI Soil Moisture Data Based on Deep Learning with an Attention Mechanism

Soil moisture (SM) is a critical variable affecting ecosystem carbon and water cycles and their feedback to climate change. In this study, we proposed a convolutional neural network (CNN) model embedded with a residual block and attention module, named SMNet, to spatially downscale the European Space Agency (ESA) Climate Change Initiative (CCI) SM product. In the SMNet model, a lightweight Convolutional Block Attention Module (CBAM) dual-attention mechanism was integrated to comprehensively extract the spatial and channel information from the high-resolution input remote sensing products, the reanalysis meteorological dataset, and the topographic data. The model was employed to downscale the ESA CCI SM from its original spatial resolution of 25 km to 1 km in California, USA, in the annual growing season (1 May to 30 September) from 2003 to 2021. The original ESA CCI SM data and in situ SM measurements (0–5 cm depth) from the International Soil Moisture Network were used to validate the model’s performance. The results show that compared with the original ESA CCI SM data, the downscaled SM data have comparable accuracy with a mean correlation (R) and root mean square error (RMSE) of 0.82 and 0.052 m3/m3, respectively. Moreover, the model generates reasonable spatiotemporal SM patterns with higher accuracy in the western region and relatively lower accuracy in the eastern Nevada mountainous area. In situ site validation results in the SCAN, the SNOTEL network, and the USCRN reveal that the R and RMSE are 0.62, 0.63, and 0.77, and 0.077 m3/m3, 0.093 m3/m3, and 0.078 m3/m3, respectively. The results are slightly lower than the validation results from the original ESA CCI SM data. Overall, the validation results suggest that the SMNet downscaling model proposed in this study has satisfactory performance in handling the task of soil moisture downscaling. The downscaled SM model not only preserves a high level of spatial consistency with the original ESA CCI SM model but also offers more intricate spatial variations in SM depending on the spatial resolution of model input data.

Comparison of Met Office regional model soil moisture with COSMOS‐UK field‐scale in situ observations

AbstractThe UK Met Office state‐of‐the‐art, deterministic, convection‐permitting, coupled land‐atmosphere, regional weather forecasting system, known as the UKV or UK Variable resolution model (Tang et al. Meteorological Applications, 2013; 20:417–426), has been operational since 2015. Science updates are regularly made to the UKV land surface data assimilation scheme when those updates improve predictions of screen temperature and humidity, since these quantities have a direct impact on atmospheric states and weather forecasts. Less attention has been paid to whether UKV soil moisture analyses are close to independent, in‐situ soil moisture observations, partly because it is difficult to make meaningful comparisons between 1.5 km2 gridded model outputs and traditional point sensor measurements. Soil moisture is recognized to be important when hydrological forecasts for runoff and rivers are required. This is because soil moisture controls the extent to which rainfall can infiltrate the soil, and the amount of surface runoff affects the timing of peak river flows (Ward & Robinson, Principles of Hydrology. McGraw‐Hill Publishing Company; 2000; Singh et al. Water Resources Research, 2021, 57, e2020WR028827). Gómez et al. (Remote Sensing, 2020; 12:3691) report benefits to river flow forecasts when using soil moisture data assimilation in the UKV system instead of a daily downscaled product from the Met Office global model. The Met Office measures soil temperature and soil moisture at Cardington (Osborne & Weedon, Journal of Hydrometeorology, 2021, 22:279–295); there is no other UK Met Office site at which soil moisture is measured. In this study, we use field‐scale (~200 m radius) soil moisture measurements from the UK Centre for Ecology and Hydrology's (UKCEH's) COSMOS‐UK network to provide independent verification and analysis of UKV soil moisture during summer 2018, an unusually dry period in the United Kingdom. We find that the match to COSMOS‐UK soil moisture observations is generally good, and that changes made to the land data assimilation approach during a recent operational upgrade had a generally beneficial impact on UKV soil moisture analyses under very dry conditions.

Drought Monitoring of Winter Wheat in Henan Province, China Based on Multi-Source Remote Sensing Data

Characterized by soil moisture content and plant growth, agricultural drought occurs when the soil moisture content is lower than the water requirement of plants. Microwave remote sensing observation has the advantages of all-weather application and sensitivity to soil moisture change. However, microwave remote sensing can only invert 0~5 cm of soil surface moisture, so it cannot effectively reflect the drought situation of farmland. Therefore, this study took Henan Province as the study area, used soil moisture active and passive (SMAP) satellite soil moisture data, employed NDVI, LST, and ET as the independent variables, and took the drought grade on the sample as the dependent variable. Using the 2017–2019 data as the training set and the 2020 data as the testing set, a random forest drought monitoring model with comprehensive influence of multiple factors was constructed based on the training set data. In the process of model training, the cross-validation method was employed to establish and verify the model. This involved allocating 80% of the sample data for model construction and reserving 20% for model verification. The results demonstrated an 85% accuracy on the training set and an 87% accuracy on the testing set. Additionally, two drought events occurring during the winter wheat growing period in Henan Province were monitored, and the validity of these droughts was confirmed using on-site soil moisture and the vegetation supply water index (VSWI). The findings indicated a high incidence of agricultural drought in the southwestern part of Henan Province, while the central and northern regions experienced a lower incidence during the jointing to heading and filling stages. Subsequently, leveraging the results from the random forest drought monitoring, this study conducted a time series analysis using the Mann–Kendall test and a spatial analysis employing Moran’s I index to examine the temporal and spatial distribution of agricultural drought in Henan Province. This analysis aimed to unveil trends in soil moisture changes affecting agricultural drought, as observed via the SMAP satellite (NASA). The results suggested a possible significant spatial auto-correlation in the occurrence of agricultural drought.

Rainfall–runoff modeling using an Adaptive Neuro-Fuzzy Inference System considering soil moisture for the Damanganga basin

ABSTRACT Rainfall is the major component of the hydrologic cycle and it is the primary source of runoff. The main purpose of this study was to estimate daily discharge by employing an Adaptive Neuro-Fuzzy Inference System (ANFIS) model using rainfall and soil moisture data at three different depths (5 cm, 100 cm and bedrock) for the Damanganga basin. The length of the data for the study period 1983–2022 is 39 years. The model employed nine membership functions for each variable of soil moisture, rainfall, discharge and 30 rules were optimized. The results were compared considering a range of model performance indicators as correlation coefficient (R2) and Nash–Sutcliffe efficiency (NSE) coefficient. The model application results shows that soil moisture at bedrock gives more precise value of daily discharge with (R2) and NSE value as 0.9936 and 0.9981, respectively, as compared to the soil moisture at depths of 5 and 100 cm. The better results obtained for the measurement of soil moisture in the deeper soil layer are consistent with the hydrological behavior anticipated for the analyzed catchment, where the root-zone soil layer is the driver of the runoff response rather than the surface observations. This study can be helpful to hydrologists in selecting appropriate rainfall–runoff models.

Using machine learning techniques to reconstruct the signal observed by the GRACE mission based on AMSR-E microwave data

Abstract This study delves into the synergy between remote sensing and satellite gravimetry, focusing on the utilization of Advanced Microwave Scanning Radiometer (AMSR-E) data for modeling delta Total Water Storage (ΔTWS) values derived from the GRACE mission. Various machine learning algorithms were employed to investigate the concordance between Gravity Recovery and Climate Experiment (GRACE) and AMSR-E observations. Despite the limited correlation in circumpolar permafrost areas, ΔTWS was successfully modeled with an accuracy of a Root Mean Square Error (RMSE) of 3.5 cm. The Amazon region exhibited a notable model error, attributed to significant ΔTWS amplitude; the overall model quality was affirmed by Normalized Root Mean Square Error (NRMSE) and Nash-Sutcliffe Efficiency (NSE) metrics. Importantly, the effectiveness of AMSR-E Soil Moisture (SM) data, encompassing C (frequency of 4–8 GHz) and X (frequency of 8–12 GHz) ranges (~0.04 m and ~0.03 m wavelength, respectively) in modeling ΔTWS, even in heavily forested equatorial regions, was demonstrated.

Water Vapor Lidar Observation and Data Assimilation for a Moist Low-Level Jet Triggering a Mesoscale Convective System

Abstract We conducted field observations using two water vapor Raman lidars (RLs) in Kyushu, Japan, to clarify the characteristics of a moist low-level jet (MLLJ), which plays a fundamental role in the formation and maintenance of mesoscale convective systems (MCSs). The two RLs observed the inside and outside of an MLLJ, providing moisture to an MCS with local heavy precipitation on 9 July 2021. Our observations revealed that the MLLJ contained large amounts of moisture below the convective mixing layer height of 1.6 km. The large amount of moisture in the MLLJ might be intensified by low-level convergences and/or water vapor buoyancy facilitated by strong horizontal wind. We conducted four data assimilation experiments: CNTL that assimilated Japan Meteorological Agency operational observation data and three other experiments that ingested the lidar-derived vertical moisture profiles as well as the operational observation data. The experiments assimilating lidar-derived vertical moisture profiles caused intensification and southwestward extensions of the low-level convergence zone, resulting in local heavy precipitation at lower latitudes in experiments assimilating lidar-derived moisture profiles than in CNTL. All three experiments ingesting vertical moisture profiles generally produced better 9-h precipitation forecasts than CNTL, implying that the assimilation of vertical moisture profiles could be well suited for numerical weather prediction of local heavy precipitation. Moreover, the experiment assimilating both of the two RL sites’ data reproduced better forecast fields than experiments assimilating a single RL site’s data, implying that data assimilation of vertical moisture profiles at multiple RL sites enables us to improve initial conditions compared to a single RL site. Significance Statement Moist low-level jets (MLLJs) are moisture-rich airflows in the low-level atmosphere that play an important role in developing mesoscale convective systems and local heavy rainfall. To better understand the mechanisms affecting the development of local heavy rainfall events and to improve our ability to forecast them, studying the moisture structures in MLLJs is important. We succeeded in observing an MLLJ in western Japan using water vapor Raman lidars (RLs), which obtained vertical moisture profiles, and revealed details of vertical moisture structures in the MLLJ. We also performed data assimilation experiments to examine the impact of assimilating vertical moisture profiles observed by the RLs. The results showed that the assimilation of the moisture data improved the forecasting of local heavy rainfall.

More than three-fold increase in compound soil and air dryness across Europe by the end of 21st century

Increases in air temperature lead to increased dryness of the air and potentially develops increased dryness in the soil. Extreme dryness (in the soil and/or in the atmosphere) affects the capacity of ecosystems for functioning and for modulating the climate. Here, we used long-term high temporal resolution (daily) soil moisture (SM) and vapor pressure deficit (VPD) data of high spatial resolution (∼0.1° × 0.1°) to show that compared to the reference period (1950–1990), the overall frequency of extreme soil dryness, extreme air dryness, and extreme compound dryness (i.e., co-occurrence of extreme soil dryness and air dryness) has increased by 1.2-fold [0.8,1.6] (median [10th,90th percentile], 1.6-fold [1,2.3], and 1.7-fold [0.9,2.5], respectively, over the last 31 years (1991–2021) across Europe. Our results also indicate that this increase in frequency of extreme compound dryness (between reference and 1991–2021 period) is largely due to increased SM-VPD coupling across Northern Europe, and due to decreasing SM and/or increasing VPD trend across Central and Mediterranean Europe. Furthermore, under the RCP8.5 (Representative Concentration Pathways 8.5) emission scenario, this increase in frequency of extreme compound dryness would be 3.3-fold [2.0,5.8], and 4.6-fold [2.3,11.9] by mid-21st century (2031–2065) and late-21st century (2066–2100), respectively. Additionally, we segregated the changes in frequency of extreme dryness across the most recent (year 2021) land cover types in Europe to show that croplands, broadleaved forest, and urban areas have experienced more than twice as much extreme dryness during 1990–2021 compared to the reference period of 1990–2021, which based on the future projection data will increase to more than three-fold by mid 21st century. Such future climate-change induced increase in extreme dryness could have negative implications for functioning of ecosystems and compromise their capacity to adapt to rapidly rising dryness levels.

A Real-Time Prediction Approach to Deep Soil Moisture Combining GNSS-R Data and a Water Movement Model in Unsaturated Soil

Deep soil moisture data have wide applications in fields such as engineering construction and agricultural production. Therefore, achieving the real-time monitoring of deep soil moisture is of significant importance. Current soil monitoring methods face challenges in conducting the large-scale, real-time monitoring of deep soil moisture. This paper innovatively proposes a real-time prediction approach to deep soil moisture combining GNSS-R data and a water movement model in unsaturated soil. This approach, built upon surface soil moisture data retrieved from GNSS-R signal inversion, integrates soil–water characteristics and soil moisture values at a depth of 1 m. By employing a deep soil moisture content prediction model, it provides predictions of soil moisture at depths from 0 to 1 m, thus realizing the large-scale, real-time dynamic monitoring of deep soil moisture. The proposed approach was validated in a study area in Goodwell, Texas County, Oklahoma, USA. Predicted values of soil moisture at a randomly selected location in the study area at depths of 0.1 m, 0.2 m, 0.5 m, and 1 m were compared with ground truth values for the period from 25 October to 19 November 2023. The results indicated that the relative error (δ) was controlled within the range of ±14%. The mean square error (MSE) ranged from 2.90 × 10−5 to 1.88 × 10−4, and the coefficient of determination (R2) ranged from 82.45% to 89.88%, indicating an overall high level of fitting between the predicted values and ground truth data. This validates the feasibility of the proposed approach, which has the potential to play a crucial role in agricultural production, geological disaster management, engineering construction, and heritage site preservation.

Enhancing agricultural automation through weather invariant soil parameter prediction using machine learning

Soil parameters are crucial aspects in increasing agricultural production. Even though Bangladesh is heavily dependent on agriculture, little research has been done regarding its automation. And a vital aspect of agricultural automation is predicting soil parameters. Generally, sensors relating to soil parameters are quite expensive and are often done in a controlled environment such as a greenhouse. However, a large scale implementation of such expensive sensors is not very feasible. This work tries to find an inexpensive solution towards predicting soil parameters such as soil moisture and temperature, both of which are crucial to the growth of crops. We focus on finding a robust relation between the above mentioned soil parameters with the nearby weather parameters such as humidity and temperature, irrespective of the weather. We apply different machine learning models like multilayer perceptron (MLP), random forest, etc. to predict the soil parameters, given the humidity and temperature of the surrounding environment. For all the experiments we have used a custom made dataset, which contains around 9000 datapoints of soil moisture & temperature, ambient humidity & temperature. The data has been collected in an uncontrolled agriculture bed via inexpensive sensors. Our results show that XGBoost regressor achieves the best results with an R2 score of 0.93 and 0.99 for soil moisture and soil temperature data respectively. This suggests very high correlation between the weather parameters and soil parameters. The model also portrayed a very low root mean squared error and mean absolute error of 0.037 & 0.015 for soil moisture and 0.001 & 0.0008 for soil temperature. Our results show that it is indeed possible to find the soil parameters from the corresponding weather, which will have great impact on mass agricultural automation. The dataset has been made publicly available at https://github.com/Nadimulhaque0403/Soil_parameter_prediction_dataset.

Joint assimilation of satellite-based surface soil moisture and vegetation conditions into the Noah-MP land surface model

This study explores the potential of integrating satellite retrievals of surface soil moisture (SSM) and vegetation conditions into the Noah-MP land surface model. In total, five data assimilation (DA) experiments were carried out. One of the experiments only assimilates SSM retrievals from the Soil Moisture Active Passive mission, two experiments only assimilate retrievals of vegetation conditions: either optical retrievals of leaf area index (LAI) from the Copernicus Global Land Service, or X-band microwave-based retrievals of vegetation optical depth (VOD) from the Advanced Microwave Scanning Radiometer 2. Additionally, two joint DA experiments are performed, each incorporating SSM and one of the vegetation products. The DA experiments are compared with a model-only run, and all experiments are evaluated using independent ground reference data of soil moisture, evapotranspiration, net ecosystem exchange and gross primary production (GPP). Assimilating only SSM improves estimates of the soil moisture profile (median SSM anomaly correlation improves with 0.02 compared to a model-only run), whereas assimilating LAI predominantly improves GPP estimates (reduction in median RMSD of 0.024 gC m−2 day−1 compared to a model-only run). The joint assimilation of SSM and vegetation conditions captures both of these improvements in a single, physically consistent analysis product. The DA increments show that this combined setup allows one satellite product to compensate for potential degradations introduced into the system by the other product. Furthermore, the joint SSM and VOD DA experiment has the smallest ensemble spread in its estimates (21% reduction in SSM spread compared to a model-only run). Overall, our results underline the potential of multi-sensor and multivariate DA, in which information from different sources is combined to improve the estimates of several land surface states and fluxes simultaneously.

Estimating soil hydraulic properties from oven-dry to full saturation using shortwave infrared imaging and inverse modeling

To minimize uncertainty related to soil processes in extreme events, we need accurate soil hydraulic properties across the entire range of soil water content. However, conventional methods are time-consuming and limited to specific ranges. To estimate soil hydraulic properties throughout the entire range, we conducted inverse modeling using upward infiltration experiments, where a shortwave infrared imaging camera was used to obtain high-resolution soil moisture data in space and time. Because the commonly used van Genuchten–Mualemmodel is unsuitable for describing soil hydraulic properties for dry conditions, we tested an alternative model, the Peters-Durner-Iden model, which considers both capillary and film water. The inverse modeling successfully estimated soil hydraulic properties for sandy loam and loam soils, and we demonstrated that the Peters-Durner-Iden model captured soil moisture dynamics better than the van Genuchten–Mualemmodel for dry conditions. However, both models could not adequately describe the soil moisture data for the other soils. The direct observation of the water flow via shortwave infrared images clarified that the reduced success was because of violating the one-dimensional flow assumption for coarse-textured soils and the micro-heterogeneity in soil hydraulic properties for soils with fine silt and clay materials.

Streamflow response to catastrophic windthrow and forest recovery in subalpine spruce forest

Windstorms account for about half of forest disturbances in Europe and are considered as drivers of natural regeneration of plant species, accompanied by significant changes in hydrological and soil moisture conditions. Our study aims to identify the effects of deforestation caused by catastrophic windthrow on streamflow characteristics in conjunction with meteorological and soil moisture data. The research was conducted in the headwater catchment (34.4 km2), in the Western Tatra Mountains, Poland. Streamflow characteristics such as average, high, and low flows, coefficient of variability, as well as flashiness of the streamflow (Richard-Baker Flashiness Index) and drought frequency (Standardised Streamflow Index) were calculated in the pre-event (2007–2013) and post-event (2013–2020) periods. The analysis was carried out for calendar months, seasons of the year, as well as for the growing and non-growing seasons. Wavelet Analysis was applied to track the changes in streamflow time series in relation to meteorological conditions (precipitation, air temperature) and remotely sensed soil wetness in the root zone. Principal Components Analysis was applied in order to identify the major drivers of the observed variability. The results revealed increasing low (+14%), average (+8%), and maximum monthly flows (16%) in the post-event period. The most significant increase in low flows (+47%, p<0.1) occurred during snowmelt. The increase in average (+31%, p<0.1) and high flows (+90%, p<0.05) occurred during mid-winter thaws. The frequency of hydrological droughts decreased by up to 36% in the post-event period, despite similar climatic conditions. The flashiness of streamflow increased during mid-winter thaws and rainfall events. The short-scale periodicity in streamflow increased in the post-event period. The results revealed the presence of 3-month periodicity in the root zone soil wetness associated with long-term soil water storage, which decreased after the catastrophic windthrow. Streamflow in the post-event period became strongly coupled with precipitation and root zone soil wetness at short period bands (∼3–16 days) in hydrologically active season. Short-scale correlation between precipitation and streamflow was decreasing with the gradual succession of vegetation.