Computers and Electronics in Agriculture, Feb 1, 2020
Detailed soil organic matter (SOM) spatial distribution maps are essential for soil management an... more Detailed soil organic matter (SOM) spatial distribution maps are essential for soil management and forestry operations. However, mapping of spatial SOM distribution over a large area is a difficult challenge, especially in regions where field samples are difficult to obtain. The objective of this research was to develop a two-stage approach to map SOM content with 10 m-resolution in Yunfu, South China with an area of 7785 km 2. In the first stage, using 10-fold cross-validation 511 artificial neural network (ANN) models were built to map SOM content based on 318 field samples from three of five sub-areas of Yunfu (ANN model area). Results indicated that the optimal ANN model with six DEM-derived variables as model inputs, i.e. ANN6, had a good model performance in ANN model area, 5.6 g/kg of root mean squared error (RMSE), 0.81 of R 2 , and 84.1% of relative overall accuracy (ROA) ± 10%, and the best generalization capability in the rest two of five sub-areas of Yunfu (extended model area), with 7.7 g/kg of RMSE, 0.58 of R 2 , and 60.7% of ROA ± 10%. In the second stage, using the reverse k-fold cross-validation extended models were developed to adapt ANN6-produced SOM content to fit field samples in the extended model areas. Results indicated the optimal extended model only required 20% of 386 field samples (5-fold) to build a stable and significant linear relationship between ANN6-produced SOM content and measured SOM content from the extended model area, and improved model accuracy with 9-21% of RMSE, 28-29% of R 2 , and 6-21% of ROA ± 10%. Thus, the two-stage method is a viable way to generate SOM content over a large area with limited number of field samples.
Ecosites are required for stand-level forest management and can be determined within a two-dimens... more Ecosites are required for stand-level forest management and can be determined within a two-dimensional edatopic grid with soil nutrient regimes (SNRs) and soil moisture regimes (SMRs) as coordinates. A new modeling method is introduced in this study to map high-resolution SNR and SMR and then to design ecosites in Nova Scotia, Canada. Using coarse-resolution soil maps and nine topo-hydrologic variables derived from high-resolution digital elevation model (DEM) data as model inputs, 511 artificial neural network (ANN) models were developed by a 10-fold cross-validation with 1507 field samples to estimate 10 m resolution SNR and SMR maps. The results showed that the optimal models for mapping SNR and SMR engaged eight and seven topo-hydrologic variables, together with three coarse-resolution soil maps, as model inputs, respectively; 82% of model-estimated SNRs were identical to field assessments, while this value was 61% for SMRs, and the produced ecosite maps had 67–68% correctness. ...
Advanced Applications for Artificial Neural Networks, 2018
High-resolution maps of soil property are considered as the most important inputs for decision su... more High-resolution maps of soil property are considered as the most important inputs for decision support and policy-making in agriculture, forestry, flood control, and environmental protection. Commonly, soil properties are mainly obtained from field surveys. Field soil surveys are generally time-consuming and expensive, with a limitation of application throughout a large area. As such, high-resolution soil property maps are only available for small areas, very often, being obtained for research purposes. In the chapter, artificial neural network (ANN) models were introduced to produce high-resolution maps of soil property. It was found that ANNs can be used to predict high-resolution soil texture, soil drainage classes, and soil organic content across landscape with reasonable accuracy and low cost. Expanding applications of the ANNs were also presented.
Soil organic carbon storage (SOCS) estimation is a crucial branch of the atmospheric–vegetation–s... more Soil organic carbon storage (SOCS) estimation is a crucial branch of the atmospheric–vegetation–soil carbon cycle study under the background of global climate change. SOCS research has increased worldwide. The objective of this study is to develop a two-stage approach with good extension capability to estimate SOCS. In the first stage, an artificial neural network (ANN) model is adopted to estimate SOCS based on 255 soil samples with five soil layers (20 cm increments to 100 cm) in Luoding, Guangdong Province, China. This method is compared with three common methods: The soil type method (STM), ordinary kriging (OK), and radial basis function (RBF) interpolation. In the second stage, a linear model is introduced to capture the regional differences and further improve the estimation accuracy of the Luoding-based ANN model when extending it to Xinxing, Guangdong Province. This is done after assessing the generalizability of the above four methods with 120 soil samples from Xinxing. Th...
The depth-specific zinc (Zn) and copper (Cu) maps with high resolution (i.e., ≤10 m) are importan... more The depth-specific zinc (Zn) and copper (Cu) maps with high resolution (i.e., ≤10 m) are important for soil and forest management and conservation. The objective of this study was to assess the effects of easily accessible model inputs, i.e., existing coarse-resolution parent material, pH, and soil texture maps with 1:1 800 000–2 800 000 scale and nine digital elevation model (DEM)-generated terrain attributes with 10 m resolution, on modelling Zn and Cu distributions of forest soil over a large area (e.g., thousands of km2). A total of 511 artificial neural network (ANN) models for each depth (20 cm increments to 100 cm) were built and evaluated by a 10-fold cross-validation with 385 soil profiles from the Yunfu forest, South China, about 4915 km2 areas. The results indicated that the optimal models for five depths engaged five to seven DEM-generated attributes together with three coarse-resolution soil attributes as inputs, respectively, and accuracies for estimating Zn and Cu var...
Cadmium (Cd) is a toxic metal and found in various soils, including forest soils. The great spati... more Cadmium (Cd) is a toxic metal and found in various soils, including forest soils. The great spatial heterogeneity in soil Cd makes it difficult to determine its distribution. Both traditional soil surveys and spatial modeling have been used to study the natural distribution of Cd. However, traditional methods are highly labor-intensive and expensive, while modeling is often encumbered by the need to select the proper predictors. In this study, based on intensive soil sampling (385 soil pits plus 64 verification soil pits) in subtropical forests in Yunfu, Guangdong, China, we examined the impacting factors and the possibility of combining existing soil information with digital elevation model (DEM)-derived variables to predict the Cd concentration at different soil depths along the landscape. A well-developed artificial neural network model (ANN), multi-variate analysis, and principal component analysis were used and compared using the same dataset. The results show that soil Cd conc...
The study on the spatial distribution of forest soil nutrients is important not only as a referen... more The study on the spatial distribution of forest soil nutrients is important not only as a reference for understanding the factors affecting soil variability, but also for the rational use of soil resources and the establishment of a virtuous cycle of forest ecosystems. The rapid development of remote sensing satellites provides an excellent opportunity to improve the accuracy of forest soil prediction models. This study aimed to explore the utility of the Gaofen-1 (GF-1) satellite in the forest soil mapping model in Luoding City, Yunfu City, Guangdong Province, Southeast China. We used 1000 m resolution coarse-resolution soil map to represent the overall regional soil nutrient status, 12.5 m resolution terrain-hydrology variables to reflect the detailed spatial distribution of soil nutrients, and 8 m resolution remote sensing variables to reflect the surface vegetation status to build terrain-hydrology artificial neural network (ANN) models and full variable ANNs, respectively. The ...
Stocks and stoichiometry of carbon (C), nitrogen (N), and phosphorus (P) in ultisols are not well... more Stocks and stoichiometry of carbon (C), nitrogen (N), and phosphorus (P) in ultisols are not well documented for converted forests. In this study, Ultisols were sampled in 175 plots from one type of secondary forest and four plantations of Masson pine (Pinus massoniana Lamb.), Slash pine (Pinus elliottii Engelm.), Eucalypt (Eucalyptus obliqua L’Hér.), and Litchi (Litchi chinensis Sonn., 1782) in Yunfu, Guangdong province, South China. Five layers of soil were sampled with a distance of 20 cm between two adjacent layers up to a depth of 100 cm. We did not find interactive effects between forest type and soil layer depth on soil organic carbon (SOC), total nitrogen (TN), and total phosphorus (TP) concentrations and storages. Storage of SOC was not different between secondary forests and Eucalypt plantations, but SOC of these two forest types were lower than that in Litchi, Masson pine, and Slash pine plantations. Soil C:P was higher in Slash pine plantations than in secondary forests....
The depth-specific soil texture map with high-resolution (i.e. ≤10 m) is essential for soil manag... more The depth-specific soil texture map with high-resolution (i.e. ≤10 m) is essential for soil management and forest silviculture. The objective of this research was to develop a modelling method to generate high-resolution soil texture maps at five depths (D1: 0-20, D2: 20-40, D3: 40-60, D4: 60-80, and D5: 80-100 cm) in Yunfu, a typical area of Udults Zone, South China. Taking a coarse-resolution soil texture (CST) map with a 1: 2,800,000 scale and nine topo-hydrologic variables derived from a digital elevation model (DEM) with 10 m-resolution as input candidates, a series of artificial neural network (ANN) models for five depths were built and evaluated by a 10fold cross-validation with 385 soil profiles from the Yunfu forest. The results indicated that the optimal model for five depths engaged five, five, five, four, and four DEM-generated variables as inputs, respectively, and model accuracies for estimating sand and clay contents varied with root mean squared error (RMSE) of 6.8-9.7%, R 2 of 0.56-0.72, and relative overall accuracy (ROA) ± 5% of 54-81%, which were better than most of other researches. An extra independent validation with 64 soil profiles outside of the model-building area also indicated that the optimal models had adequate capabilities for generalization with RMSE of 9.2-12.2%, R 2 of 0.33-0.47, and ROA ± 5% of 37-53%. The depth-specific sand and clay content maps with 10 m-resolution generated from the optimal models in Yunfu showed more detailed information than the CST map, and could reflected the influence of the DEM-derived topo-hydrologic variables. Based on the generated maps, horizontal characteristics of soil texture in the study area exhibited an obvious process of clay translocation from the topsoil (D1) to subsoil (D2-5), a maximum accumulation of clay in D4, and a dominant sandy soil in the topsoil (D1). Thus, the modelling method, i.e. developing ANNs with k-fold cross-validation, can be used to generate depth-specific soil texture maps in Udults Zone, South China. In addition, the generated high-resolution maps can clearly show the changes of soil texture in three-dimension.
The noises of remote sensing images, caused by imaging system and ground environment, negatively ... more The noises of remote sensing images, caused by imaging system and ground environment, negatively affect the accuracy and efficiency in extracting forest information from remote sensing images. The denoising is critical for image classifications for forest areas. The objective of this research is to assess the effectiveness of currently used spatial filtering methods for extracting with forest information related from Landsat 5 TM images. Five spatial filtering methods including low-pass filter, median filter, mean filter, sigma filter and enhanced self-adaptive filter were examined. A set of evaluation indices was designed to assess the ability of each denoising method for flatness, edge/boundary retention and enhancement. Based on the designed evaluation indices and visual assessment, it was found that sigma filter (D=1) and enhanced self-adaptive filter were the most effective denoising methods in classifying TM images for forest areas.
Forest ecosite reflects the local site conditions that are meaningful to forest productivity as w... more Forest ecosite reflects the local site conditions that are meaningful to forest productivity as well as basic ecological functions. Field assessments of vegetation and soil types are often used to identify forest ecosites. However, the production of high-resolution ecosite maps for large areas from interpolating field data is difficult because of high spatial variation and associated costs and time requirements. Indices of soil moisture and nutrient regimes (i.e., SMR and SNR) introduced in this study reflect the combined effects of biogeochemical and topographic factors on forest growth. The objective of this research is to present a method for creating high-resolution forest ecosite maps based on computergenerated predictions of SMR and SNR for an area in Atlantic Canada covering about 4.3 × 10 6 hectares (ha) of forestland. Field data from 1,507 forest ecosystem classification plots were used to assess the accuracy of the ecosite maps produced. Using model predictions of SMR and SNR alone, ecosite maps were 61 and 59% correct in identifying 10 Acadian-and Maritime-Boreal-region ecosite types, respectively. This method provides an operational framework for the production of high-resolution maps of forest ecosites over large areas without the need for data from expensive, supplementary field surveys. Ecosites, as stand-level units in ecological land classification systems, describe a suite of site conditions that characterize forest productivity. They also provide an ecological basis for grouping vegetation and soil types 1. High-resolution ecosite maps (≤10 m) are useful for forest stand-level planning purposes, such as growth and yield analysis, best management practices implementation, and forest ecosystem management 2-4. These maps provide forest managers, conservationists, and governmental organizations the ability to develop silvicultural systems, forest management plans, and environment-protection protocol and policy. The principal method in identifying ecosites is based on gaging soil and vegetation types identified from a number of easily observable field indicators 5-8. However, field procedures can be subjective in view of the fact that ground vegetation changes seasonally over the short-term and with forest stand succession over the longer term 9-11. Also, from a mapping point of view, generating ecosite maps from the interpolation of point assessments would require many field surveys be carried out to produce maps of acceptable detail (i.e., resolution + accuracy), given the inherent complexity of forest landscapes. In general, field data-collection procedures associated with ecosite mapping are time consuming and expensive to use, particularly in large areas spanning more than ten to hundreds of thousands of hectares (ha). In the past two decades, new methods have been developed, largely based on air-photo-and model-based interpretation, with an aim to increase map production rates and reduce costs 12-14. For example, MacMillan et al. 14 reported that by mapping ecosites with automated-predictive-mapping procedures, total mapping costs for a 3 × 10 6 ha forest, processed at a 25-m resolution, were reduced from more than
Biome‐specific soil respiration (Rs) has important yet different roles in both the carbon cycle a... more Biome‐specific soil respiration (Rs) has important yet different roles in both the carbon cycle and climate change from regional to global scales. To date, no comparable studies related to global biome‐specific Rs have been conducted applying comprehensive global Rs databases. The goal of this study was to develop artificial neural network (ANN) models capable of spatially estimating global Rs and to evaluate the effects of interannual climate variations on 10 major biomes. We used 1976 annual Rs field records extracted from global Rs literature to train and test the ANN models. We determined that the best ANN model for predicting biome‐specific global annual Rs was the one that applied mean annual temperature (MAT), mean annual precipitation (MAP), and biome type as inputs (r2 = 0.60). The ANN models reported an average global Rs of 93.3 ± 6.1 Pg C yr−1 from 1960 to 2012 and an increasing trend in average global annual Rs of 0.04 Pg C yr−1. Estimated annual Rs increased with increases in MAT and MAP in cropland, boreal forest, grassland, shrubland, and wetland biomes. Additionally, estimated annual Rs decreased with increases in MAT and increased with increases in MAP in desert and tundra biomes, and only significantly decreased with increases in MAT (r2 = 0.87) in the savannah biome. The developed biome‐specific global Rs database for global land and soil carbon models will aid in understanding the mechanisms underlying variations in soil carbon dynamics and in quantifying uncertainty in the global soil carbon cycle.
Soil property maps are considered as the most important input information for decision support an... more Soil property maps are considered as the most important input information for decision support and policy making in agriculture, forestry, flood control as well as environmental protection. Traditionally, soil property maps are mainly obtained from field surveys. Field soil survey is generally time consuming and expensive, which limited it application over a large area. As such, high resolution soil property maps are only available for small areas, very often, being obtained for research purposes. In this research, artificial neural network technology was used to generate high resolutions soil property maps. Hydrological parameters derived from digital elevation maps combined with information extracted from existing coarse resolution soil maps were used as input for the proposed model. Detailed soil survey information from Black Brook Watershed in Northern New Brunswick was used to test the model performance. We found that ANN models base model can be used to predict soil texture, s...
Exceedance of water-quality standards is important in assessing water quality. The effectiveness ... more Exceedance of water-quality standards is important in assessing water quality. The effectiveness of soil conservation Beneficial Management Practices (BMPs) should be measured according to the BMPs' impact on exceedance frequencies. However, estimating exceedance frequencies for different management scenarios with field measurements is practically impossible due to difficulties in obtaining adequate data for analysing different combinations of BMPs. The objective of this modeling research was to analyse exceedance frequencies for different management strategies applied in the Black Brook Watershed (BBW). Daily concentrations of total suspended sediments (TSS) and soluble phosphorous (sol-P) were predicted with the Soil and Water Assessment Tool (SWAT) and assessed against water-quality standards from the Canadian Council of Ministers of the Environment (CCME) and National Agri-Environmental Standards Initiative-Ideal Performance Standards (NAESI-IPS). The investigated BMPs inclu...
h i g h l i g h t s Mean annual soil respiration rate was 33.65 t CO 2 ha À1 year À1 across Chine... more h i g h l i g h t s Mean annual soil respiration rate was 33.65 t CO 2 ha À1 year À1 across Chinese forest ecosystems. Mean Q 10 value of 1.28 was lower than the world average (1.4e2.0). Artificial neural network model may effectively predict Rs across Chinese forest ecosystems. Q 10 values derived from the soil temperature significantly increased with elevation and latitude.
International Geoscience and Remote Sensing Symposium (IGARSS), 2006
Improving vegetation classification accuracy of remote sensing images has an important significat... more Improving vegetation classification accuracy of remote sensing images has an important signification to apply remote sensing in forest areas. In this study, taking the Landsat 5 TM images of the forest areas in northern New Brunswick, Canada as an example, potential solar radiation data and slope position data are introduced to classifying forest land. The objective is to examine the ability of increasing the vegetation classification by the auxiliary data. After assessing the accuracy by using error the matrix method, the results show that potential solar radiation data and slope position data can obviously improve classification accuracy of TM images in forest areas.
Ying yong sheng tai xue bao = The journal of applied ecology / Zhongguo sheng tai xue xue hui, Zhongguo ke xue yuan Shenyang ying yong sheng tai yan jiu suo zhu ban, 2006
To improve the accuracy of automatic classification and identification of TM remote sensing image... more To improve the accuracy of automatic classification and identification of TM remote sensing images in forest area, an expert system for automatically classifying and identifying deciduous-conifer mixed forest was built up, based on the GIS technique, quantitative analysis on the internal relations between geographic factors such as DEM and slope aspect and environment factors like soil type, and qualitative analysis on the spectrum information and preclassification information of sensing images, aimed to build a classification knowledge system. Taking the TM remote sensing image of Wangqing Forest Bureau in Jilin Province as an example, the study showed that this expert system could obviously reduce the influence of mixed pixel and terrain shadow. The classification precision of this system was increased by 14.22%, compared with that of Iterative Self-Organizing Data Analysis Techniques Algorithm (ISODATA) unsupervised classification, and the Kappa index was 0.7556, which could help...
Digital elevation model (DEM) is often used for hydrologic modeling, land use planning, engineeri... more Digital elevation model (DEM) is often used for hydrologic modeling, land use planning, engineering design and environmental protection. Research is required to assess the need of updating existing conventional DEM using higher resolution and more accurate DEMs, including light detection and ranging (LiDAR) DEM. The objective of this study was to evaluate effects of DEM accuracy and resolution on hydrologic parameters and modeling in an agriculture-dominated watershed. DEMs compared included 1 m and 10 m LiDAR based DEMs, and a conventional 10 m DEM obtained with aerial photogrammetry method. Hydrologic parameters assessed included elevation, sub-basin area and boundaries, drainage networks, slope and slope length. DEM derived hydrological parameters were used to estimate soil loss in Black Brook Watershed, New Brunswick using Revised Universal Soil Loss Equation (RUSLE). Results indicated that DEM resolution had substantial influence on the sub-basins boundaries, sub-basin area, and distribution of water flow lines. Field investigation confirmed that most of the water flow lines derived from 1 m LiDAR based DEM were accurate and a number of flow diversion terraces (FDT) failures had been identified with help of LiDAR 1 m DEM. Both
Soil conservation benefi cial management practices (BMPs) are eff ective at controlling soil loss... more Soil conservation benefi cial management practices (BMPs) are eff ective at controlling soil loss from farmlands and minimizing water pollution in agricultural watersheds. However, costs associated with implementing and maintaining these practices are high and often deter farmers from using them. Consequently, it is necessary to conduct cost-benefi t analysis of BMP implementation to assist decision-makers with planning to provide the greatest level of environmental protection with limited resources and funding. Th e Soil and Water Assessment Tool (SWAT) was used to evaluate the effi cacy of fl ow diversion terraces (FDT) in abating sediment yield at the outlet of Black Brook Watershed (BBW), northwestern New Brunswick. Diff erent FDT-implementation scenarios were expressed as the ratio of land area protected by FDT to the total cultivated area. From this analysis, we found that average annual sediment yield decreased exponentially with increased FDT protection. When the proportion of FDT-protected areas was low, sediment reductions caused by FDT increased sharply with increasing use of FDT. Similarly, marginal sediment yield abatement costs (dollar per tonne of sediment reduction) increased exponentially with increasing proportion of FDT-protected area. Th e results indicated that increasing land protection with FDT from 6 to 50% would result in a reduction of about 2.1 tonne ha-1 yr-1
Slope is a metric that is essential to describe surface hydrological processes, including overlan... more Slope is a metric that is essential to describe surface hydrological processes, including overland flow, soil erosion, and sediment transport. Most commercial GIS have built-in functions to calculate the slope from Digital Elevation Models (DEMs) by means of average neighbourhood methods that are appropriate for coarse-resolution DEMs. Emergence of high-resolution DEMs from LiDAR data creates a need to reassess the suitability of existing algorithms for calculating slope in hydrological applications. In this study, we investigate the properties of two different slope-calculation methods: an average-neighbourhood-slope (ANS) and a downhill-slope (DHS) method. Conceptually, the DHS method provides a more intuitive description of surface water-flow characteristics in an uneven terrain. DEMs of five different types were used to evaluate the methods, namely a 1-m and 10-m resolution DEM interpolated from irregular elevation point-data generated with conventional photogrammetric techniques, and a 1-m, 5-m, and 10-m resolution DEM derived from LiDAR data. The slopes calculated were summarized for the entire watershed, along mapped streams, and within pre-defined 'stream buffers'. Slopes generated for the entire watershed with 1-m resolution LiDAR DEM indicated that the ANS method on an average produced smaller slopes than the DHS method (0Ð64°). A similar trend was observed in stream buffers, with greatest slope differences (S) between methods within 20-m buffers, when the 1-m LiDAR-based DEM was used (S D 1Ð12°). In contrast, the ANS-calculated slopes along mapped streams were generally larger than those calculated with the DHS method for LiDAR-based DEMs (S D 0Ð81°). The results from this study signal the need for caution when estimating slopes along streams from high-accuracy, LiDAR-generated DEMs.
Computers and Electronics in Agriculture, Feb 1, 2020
Detailed soil organic matter (SOM) spatial distribution maps are essential for soil management an... more Detailed soil organic matter (SOM) spatial distribution maps are essential for soil management and forestry operations. However, mapping of spatial SOM distribution over a large area is a difficult challenge, especially in regions where field samples are difficult to obtain. The objective of this research was to develop a two-stage approach to map SOM content with 10 m-resolution in Yunfu, South China with an area of 7785 km 2. In the first stage, using 10-fold cross-validation 511 artificial neural network (ANN) models were built to map SOM content based on 318 field samples from three of five sub-areas of Yunfu (ANN model area). Results indicated that the optimal ANN model with six DEM-derived variables as model inputs, i.e. ANN6, had a good model performance in ANN model area, 5.6 g/kg of root mean squared error (RMSE), 0.81 of R 2 , and 84.1% of relative overall accuracy (ROA) ± 10%, and the best generalization capability in the rest two of five sub-areas of Yunfu (extended model area), with 7.7 g/kg of RMSE, 0.58 of R 2 , and 60.7% of ROA ± 10%. In the second stage, using the reverse k-fold cross-validation extended models were developed to adapt ANN6-produced SOM content to fit field samples in the extended model areas. Results indicated the optimal extended model only required 20% of 386 field samples (5-fold) to build a stable and significant linear relationship between ANN6-produced SOM content and measured SOM content from the extended model area, and improved model accuracy with 9-21% of RMSE, 28-29% of R 2 , and 6-21% of ROA ± 10%. Thus, the two-stage method is a viable way to generate SOM content over a large area with limited number of field samples.
Ecosites are required for stand-level forest management and can be determined within a two-dimens... more Ecosites are required for stand-level forest management and can be determined within a two-dimensional edatopic grid with soil nutrient regimes (SNRs) and soil moisture regimes (SMRs) as coordinates. A new modeling method is introduced in this study to map high-resolution SNR and SMR and then to design ecosites in Nova Scotia, Canada. Using coarse-resolution soil maps and nine topo-hydrologic variables derived from high-resolution digital elevation model (DEM) data as model inputs, 511 artificial neural network (ANN) models were developed by a 10-fold cross-validation with 1507 field samples to estimate 10 m resolution SNR and SMR maps. The results showed that the optimal models for mapping SNR and SMR engaged eight and seven topo-hydrologic variables, together with three coarse-resolution soil maps, as model inputs, respectively; 82% of model-estimated SNRs were identical to field assessments, while this value was 61% for SMRs, and the produced ecosite maps had 67–68% correctness. ...
Advanced Applications for Artificial Neural Networks, 2018
High-resolution maps of soil property are considered as the most important inputs for decision su... more High-resolution maps of soil property are considered as the most important inputs for decision support and policy-making in agriculture, forestry, flood control, and environmental protection. Commonly, soil properties are mainly obtained from field surveys. Field soil surveys are generally time-consuming and expensive, with a limitation of application throughout a large area. As such, high-resolution soil property maps are only available for small areas, very often, being obtained for research purposes. In the chapter, artificial neural network (ANN) models were introduced to produce high-resolution maps of soil property. It was found that ANNs can be used to predict high-resolution soil texture, soil drainage classes, and soil organic content across landscape with reasonable accuracy and low cost. Expanding applications of the ANNs were also presented.
Soil organic carbon storage (SOCS) estimation is a crucial branch of the atmospheric–vegetation–s... more Soil organic carbon storage (SOCS) estimation is a crucial branch of the atmospheric–vegetation–soil carbon cycle study under the background of global climate change. SOCS research has increased worldwide. The objective of this study is to develop a two-stage approach with good extension capability to estimate SOCS. In the first stage, an artificial neural network (ANN) model is adopted to estimate SOCS based on 255 soil samples with five soil layers (20 cm increments to 100 cm) in Luoding, Guangdong Province, China. This method is compared with three common methods: The soil type method (STM), ordinary kriging (OK), and radial basis function (RBF) interpolation. In the second stage, a linear model is introduced to capture the regional differences and further improve the estimation accuracy of the Luoding-based ANN model when extending it to Xinxing, Guangdong Province. This is done after assessing the generalizability of the above four methods with 120 soil samples from Xinxing. Th...
The depth-specific zinc (Zn) and copper (Cu) maps with high resolution (i.e., ≤10 m) are importan... more The depth-specific zinc (Zn) and copper (Cu) maps with high resolution (i.e., ≤10 m) are important for soil and forest management and conservation. The objective of this study was to assess the effects of easily accessible model inputs, i.e., existing coarse-resolution parent material, pH, and soil texture maps with 1:1 800 000–2 800 000 scale and nine digital elevation model (DEM)-generated terrain attributes with 10 m resolution, on modelling Zn and Cu distributions of forest soil over a large area (e.g., thousands of km2). A total of 511 artificial neural network (ANN) models for each depth (20 cm increments to 100 cm) were built and evaluated by a 10-fold cross-validation with 385 soil profiles from the Yunfu forest, South China, about 4915 km2 areas. The results indicated that the optimal models for five depths engaged five to seven DEM-generated attributes together with three coarse-resolution soil attributes as inputs, respectively, and accuracies for estimating Zn and Cu var...
Cadmium (Cd) is a toxic metal and found in various soils, including forest soils. The great spati... more Cadmium (Cd) is a toxic metal and found in various soils, including forest soils. The great spatial heterogeneity in soil Cd makes it difficult to determine its distribution. Both traditional soil surveys and spatial modeling have been used to study the natural distribution of Cd. However, traditional methods are highly labor-intensive and expensive, while modeling is often encumbered by the need to select the proper predictors. In this study, based on intensive soil sampling (385 soil pits plus 64 verification soil pits) in subtropical forests in Yunfu, Guangdong, China, we examined the impacting factors and the possibility of combining existing soil information with digital elevation model (DEM)-derived variables to predict the Cd concentration at different soil depths along the landscape. A well-developed artificial neural network model (ANN), multi-variate analysis, and principal component analysis were used and compared using the same dataset. The results show that soil Cd conc...
The study on the spatial distribution of forest soil nutrients is important not only as a referen... more The study on the spatial distribution of forest soil nutrients is important not only as a reference for understanding the factors affecting soil variability, but also for the rational use of soil resources and the establishment of a virtuous cycle of forest ecosystems. The rapid development of remote sensing satellites provides an excellent opportunity to improve the accuracy of forest soil prediction models. This study aimed to explore the utility of the Gaofen-1 (GF-1) satellite in the forest soil mapping model in Luoding City, Yunfu City, Guangdong Province, Southeast China. We used 1000 m resolution coarse-resolution soil map to represent the overall regional soil nutrient status, 12.5 m resolution terrain-hydrology variables to reflect the detailed spatial distribution of soil nutrients, and 8 m resolution remote sensing variables to reflect the surface vegetation status to build terrain-hydrology artificial neural network (ANN) models and full variable ANNs, respectively. The ...
Stocks and stoichiometry of carbon (C), nitrogen (N), and phosphorus (P) in ultisols are not well... more Stocks and stoichiometry of carbon (C), nitrogen (N), and phosphorus (P) in ultisols are not well documented for converted forests. In this study, Ultisols were sampled in 175 plots from one type of secondary forest and four plantations of Masson pine (Pinus massoniana Lamb.), Slash pine (Pinus elliottii Engelm.), Eucalypt (Eucalyptus obliqua L’Hér.), and Litchi (Litchi chinensis Sonn., 1782) in Yunfu, Guangdong province, South China. Five layers of soil were sampled with a distance of 20 cm between two adjacent layers up to a depth of 100 cm. We did not find interactive effects between forest type and soil layer depth on soil organic carbon (SOC), total nitrogen (TN), and total phosphorus (TP) concentrations and storages. Storage of SOC was not different between secondary forests and Eucalypt plantations, but SOC of these two forest types were lower than that in Litchi, Masson pine, and Slash pine plantations. Soil C:P was higher in Slash pine plantations than in secondary forests....
The depth-specific soil texture map with high-resolution (i.e. ≤10 m) is essential for soil manag... more The depth-specific soil texture map with high-resolution (i.e. ≤10 m) is essential for soil management and forest silviculture. The objective of this research was to develop a modelling method to generate high-resolution soil texture maps at five depths (D1: 0-20, D2: 20-40, D3: 40-60, D4: 60-80, and D5: 80-100 cm) in Yunfu, a typical area of Udults Zone, South China. Taking a coarse-resolution soil texture (CST) map with a 1: 2,800,000 scale and nine topo-hydrologic variables derived from a digital elevation model (DEM) with 10 m-resolution as input candidates, a series of artificial neural network (ANN) models for five depths were built and evaluated by a 10fold cross-validation with 385 soil profiles from the Yunfu forest. The results indicated that the optimal model for five depths engaged five, five, five, four, and four DEM-generated variables as inputs, respectively, and model accuracies for estimating sand and clay contents varied with root mean squared error (RMSE) of 6.8-9.7%, R 2 of 0.56-0.72, and relative overall accuracy (ROA) ± 5% of 54-81%, which were better than most of other researches. An extra independent validation with 64 soil profiles outside of the model-building area also indicated that the optimal models had adequate capabilities for generalization with RMSE of 9.2-12.2%, R 2 of 0.33-0.47, and ROA ± 5% of 37-53%. The depth-specific sand and clay content maps with 10 m-resolution generated from the optimal models in Yunfu showed more detailed information than the CST map, and could reflected the influence of the DEM-derived topo-hydrologic variables. Based on the generated maps, horizontal characteristics of soil texture in the study area exhibited an obvious process of clay translocation from the topsoil (D1) to subsoil (D2-5), a maximum accumulation of clay in D4, and a dominant sandy soil in the topsoil (D1). Thus, the modelling method, i.e. developing ANNs with k-fold cross-validation, can be used to generate depth-specific soil texture maps in Udults Zone, South China. In addition, the generated high-resolution maps can clearly show the changes of soil texture in three-dimension.
The noises of remote sensing images, caused by imaging system and ground environment, negatively ... more The noises of remote sensing images, caused by imaging system and ground environment, negatively affect the accuracy and efficiency in extracting forest information from remote sensing images. The denoising is critical for image classifications for forest areas. The objective of this research is to assess the effectiveness of currently used spatial filtering methods for extracting with forest information related from Landsat 5 TM images. Five spatial filtering methods including low-pass filter, median filter, mean filter, sigma filter and enhanced self-adaptive filter were examined. A set of evaluation indices was designed to assess the ability of each denoising method for flatness, edge/boundary retention and enhancement. Based on the designed evaluation indices and visual assessment, it was found that sigma filter (D=1) and enhanced self-adaptive filter were the most effective denoising methods in classifying TM images for forest areas.
Forest ecosite reflects the local site conditions that are meaningful to forest productivity as w... more Forest ecosite reflects the local site conditions that are meaningful to forest productivity as well as basic ecological functions. Field assessments of vegetation and soil types are often used to identify forest ecosites. However, the production of high-resolution ecosite maps for large areas from interpolating field data is difficult because of high spatial variation and associated costs and time requirements. Indices of soil moisture and nutrient regimes (i.e., SMR and SNR) introduced in this study reflect the combined effects of biogeochemical and topographic factors on forest growth. The objective of this research is to present a method for creating high-resolution forest ecosite maps based on computergenerated predictions of SMR and SNR for an area in Atlantic Canada covering about 4.3 × 10 6 hectares (ha) of forestland. Field data from 1,507 forest ecosystem classification plots were used to assess the accuracy of the ecosite maps produced. Using model predictions of SMR and SNR alone, ecosite maps were 61 and 59% correct in identifying 10 Acadian-and Maritime-Boreal-region ecosite types, respectively. This method provides an operational framework for the production of high-resolution maps of forest ecosites over large areas without the need for data from expensive, supplementary field surveys. Ecosites, as stand-level units in ecological land classification systems, describe a suite of site conditions that characterize forest productivity. They also provide an ecological basis for grouping vegetation and soil types 1. High-resolution ecosite maps (≤10 m) are useful for forest stand-level planning purposes, such as growth and yield analysis, best management practices implementation, and forest ecosystem management 2-4. These maps provide forest managers, conservationists, and governmental organizations the ability to develop silvicultural systems, forest management plans, and environment-protection protocol and policy. The principal method in identifying ecosites is based on gaging soil and vegetation types identified from a number of easily observable field indicators 5-8. However, field procedures can be subjective in view of the fact that ground vegetation changes seasonally over the short-term and with forest stand succession over the longer term 9-11. Also, from a mapping point of view, generating ecosite maps from the interpolation of point assessments would require many field surveys be carried out to produce maps of acceptable detail (i.e., resolution + accuracy), given the inherent complexity of forest landscapes. In general, field data-collection procedures associated with ecosite mapping are time consuming and expensive to use, particularly in large areas spanning more than ten to hundreds of thousands of hectares (ha). In the past two decades, new methods have been developed, largely based on air-photo-and model-based interpretation, with an aim to increase map production rates and reduce costs 12-14. For example, MacMillan et al. 14 reported that by mapping ecosites with automated-predictive-mapping procedures, total mapping costs for a 3 × 10 6 ha forest, processed at a 25-m resolution, were reduced from more than
Biome‐specific soil respiration (Rs) has important yet different roles in both the carbon cycle a... more Biome‐specific soil respiration (Rs) has important yet different roles in both the carbon cycle and climate change from regional to global scales. To date, no comparable studies related to global biome‐specific Rs have been conducted applying comprehensive global Rs databases. The goal of this study was to develop artificial neural network (ANN) models capable of spatially estimating global Rs and to evaluate the effects of interannual climate variations on 10 major biomes. We used 1976 annual Rs field records extracted from global Rs literature to train and test the ANN models. We determined that the best ANN model for predicting biome‐specific global annual Rs was the one that applied mean annual temperature (MAT), mean annual precipitation (MAP), and biome type as inputs (r2 = 0.60). The ANN models reported an average global Rs of 93.3 ± 6.1 Pg C yr−1 from 1960 to 2012 and an increasing trend in average global annual Rs of 0.04 Pg C yr−1. Estimated annual Rs increased with increases in MAT and MAP in cropland, boreal forest, grassland, shrubland, and wetland biomes. Additionally, estimated annual Rs decreased with increases in MAT and increased with increases in MAP in desert and tundra biomes, and only significantly decreased with increases in MAT (r2 = 0.87) in the savannah biome. The developed biome‐specific global Rs database for global land and soil carbon models will aid in understanding the mechanisms underlying variations in soil carbon dynamics and in quantifying uncertainty in the global soil carbon cycle.
Soil property maps are considered as the most important input information for decision support an... more Soil property maps are considered as the most important input information for decision support and policy making in agriculture, forestry, flood control as well as environmental protection. Traditionally, soil property maps are mainly obtained from field surveys. Field soil survey is generally time consuming and expensive, which limited it application over a large area. As such, high resolution soil property maps are only available for small areas, very often, being obtained for research purposes. In this research, artificial neural network technology was used to generate high resolutions soil property maps. Hydrological parameters derived from digital elevation maps combined with information extracted from existing coarse resolution soil maps were used as input for the proposed model. Detailed soil survey information from Black Brook Watershed in Northern New Brunswick was used to test the model performance. We found that ANN models base model can be used to predict soil texture, s...
Exceedance of water-quality standards is important in assessing water quality. The effectiveness ... more Exceedance of water-quality standards is important in assessing water quality. The effectiveness of soil conservation Beneficial Management Practices (BMPs) should be measured according to the BMPs' impact on exceedance frequencies. However, estimating exceedance frequencies for different management scenarios with field measurements is practically impossible due to difficulties in obtaining adequate data for analysing different combinations of BMPs. The objective of this modeling research was to analyse exceedance frequencies for different management strategies applied in the Black Brook Watershed (BBW). Daily concentrations of total suspended sediments (TSS) and soluble phosphorous (sol-P) were predicted with the Soil and Water Assessment Tool (SWAT) and assessed against water-quality standards from the Canadian Council of Ministers of the Environment (CCME) and National Agri-Environmental Standards Initiative-Ideal Performance Standards (NAESI-IPS). The investigated BMPs inclu...
h i g h l i g h t s Mean annual soil respiration rate was 33.65 t CO 2 ha À1 year À1 across Chine... more h i g h l i g h t s Mean annual soil respiration rate was 33.65 t CO 2 ha À1 year À1 across Chinese forest ecosystems. Mean Q 10 value of 1.28 was lower than the world average (1.4e2.0). Artificial neural network model may effectively predict Rs across Chinese forest ecosystems. Q 10 values derived from the soil temperature significantly increased with elevation and latitude.
International Geoscience and Remote Sensing Symposium (IGARSS), 2006
Improving vegetation classification accuracy of remote sensing images has an important significat... more Improving vegetation classification accuracy of remote sensing images has an important signification to apply remote sensing in forest areas. In this study, taking the Landsat 5 TM images of the forest areas in northern New Brunswick, Canada as an example, potential solar radiation data and slope position data are introduced to classifying forest land. The objective is to examine the ability of increasing the vegetation classification by the auxiliary data. After assessing the accuracy by using error the matrix method, the results show that potential solar radiation data and slope position data can obviously improve classification accuracy of TM images in forest areas.
Ying yong sheng tai xue bao = The journal of applied ecology / Zhongguo sheng tai xue xue hui, Zhongguo ke xue yuan Shenyang ying yong sheng tai yan jiu suo zhu ban, 2006
To improve the accuracy of automatic classification and identification of TM remote sensing image... more To improve the accuracy of automatic classification and identification of TM remote sensing images in forest area, an expert system for automatically classifying and identifying deciduous-conifer mixed forest was built up, based on the GIS technique, quantitative analysis on the internal relations between geographic factors such as DEM and slope aspect and environment factors like soil type, and qualitative analysis on the spectrum information and preclassification information of sensing images, aimed to build a classification knowledge system. Taking the TM remote sensing image of Wangqing Forest Bureau in Jilin Province as an example, the study showed that this expert system could obviously reduce the influence of mixed pixel and terrain shadow. The classification precision of this system was increased by 14.22%, compared with that of Iterative Self-Organizing Data Analysis Techniques Algorithm (ISODATA) unsupervised classification, and the Kappa index was 0.7556, which could help...
Digital elevation model (DEM) is often used for hydrologic modeling, land use planning, engineeri... more Digital elevation model (DEM) is often used for hydrologic modeling, land use planning, engineering design and environmental protection. Research is required to assess the need of updating existing conventional DEM using higher resolution and more accurate DEMs, including light detection and ranging (LiDAR) DEM. The objective of this study was to evaluate effects of DEM accuracy and resolution on hydrologic parameters and modeling in an agriculture-dominated watershed. DEMs compared included 1 m and 10 m LiDAR based DEMs, and a conventional 10 m DEM obtained with aerial photogrammetry method. Hydrologic parameters assessed included elevation, sub-basin area and boundaries, drainage networks, slope and slope length. DEM derived hydrological parameters were used to estimate soil loss in Black Brook Watershed, New Brunswick using Revised Universal Soil Loss Equation (RUSLE). Results indicated that DEM resolution had substantial influence on the sub-basins boundaries, sub-basin area, and distribution of water flow lines. Field investigation confirmed that most of the water flow lines derived from 1 m LiDAR based DEM were accurate and a number of flow diversion terraces (FDT) failures had been identified with help of LiDAR 1 m DEM. Both
Soil conservation benefi cial management practices (BMPs) are eff ective at controlling soil loss... more Soil conservation benefi cial management practices (BMPs) are eff ective at controlling soil loss from farmlands and minimizing water pollution in agricultural watersheds. However, costs associated with implementing and maintaining these practices are high and often deter farmers from using them. Consequently, it is necessary to conduct cost-benefi t analysis of BMP implementation to assist decision-makers with planning to provide the greatest level of environmental protection with limited resources and funding. Th e Soil and Water Assessment Tool (SWAT) was used to evaluate the effi cacy of fl ow diversion terraces (FDT) in abating sediment yield at the outlet of Black Brook Watershed (BBW), northwestern New Brunswick. Diff erent FDT-implementation scenarios were expressed as the ratio of land area protected by FDT to the total cultivated area. From this analysis, we found that average annual sediment yield decreased exponentially with increased FDT protection. When the proportion of FDT-protected areas was low, sediment reductions caused by FDT increased sharply with increasing use of FDT. Similarly, marginal sediment yield abatement costs (dollar per tonne of sediment reduction) increased exponentially with increasing proportion of FDT-protected area. Th e results indicated that increasing land protection with FDT from 6 to 50% would result in a reduction of about 2.1 tonne ha-1 yr-1
Slope is a metric that is essential to describe surface hydrological processes, including overlan... more Slope is a metric that is essential to describe surface hydrological processes, including overland flow, soil erosion, and sediment transport. Most commercial GIS have built-in functions to calculate the slope from Digital Elevation Models (DEMs) by means of average neighbourhood methods that are appropriate for coarse-resolution DEMs. Emergence of high-resolution DEMs from LiDAR data creates a need to reassess the suitability of existing algorithms for calculating slope in hydrological applications. In this study, we investigate the properties of two different slope-calculation methods: an average-neighbourhood-slope (ANS) and a downhill-slope (DHS) method. Conceptually, the DHS method provides a more intuitive description of surface water-flow characteristics in an uneven terrain. DEMs of five different types were used to evaluate the methods, namely a 1-m and 10-m resolution DEM interpolated from irregular elevation point-data generated with conventional photogrammetric techniques, and a 1-m, 5-m, and 10-m resolution DEM derived from LiDAR data. The slopes calculated were summarized for the entire watershed, along mapped streams, and within pre-defined 'stream buffers'. Slopes generated for the entire watershed with 1-m resolution LiDAR DEM indicated that the ANS method on an average produced smaller slopes than the DHS method (0Ð64°). A similar trend was observed in stream buffers, with greatest slope differences (S) between methods within 20-m buffers, when the 1-m LiDAR-based DEM was used (S D 1Ð12°). In contrast, the ANS-calculated slopes along mapped streams were generally larger than those calculated with the DHS method for LiDAR-based DEMs (S D 0Ð81°). The results from this study signal the need for caution when estimating slopes along streams from high-accuracy, LiDAR-generated DEMs.
Uploads
Papers by zhengyong zhao