Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

World's Largest Regreening Project Promotes Healthy Tree Growth and Nutrient Accumulation Up to 40-Years Post Restoration

World's Largest Regreening Project Promotes Healthy Tree Growth and Nutrient Accumulation Up to... 17 Mining and smelting degraded landscapes are characterised by heavily eroded, acidic soils that are 18 contaminated with toxic metals and depleted of essential nutrients. Restoring forests on these 19 landscapes has been highlighted to support carbon (C) mitigation measures and protect biodiversity. 20 Understanding how tree growth and aboveground nutrient accumulation changes following restoration 21 will be essential to planning future forest restoration projects. In this study, we assessed aboveground 22 biomass (AGB) and aboveground nutrient (calcium (Ca), magnesium (Mg), nitrogen (N), phosphorous (P), 23 potassium (K), and C) pools across as series of sites ranging in age from 15 to 42 years old; to determine 24 the effects of erosion on AGB and AGB nutrient pools each site was categorized as “stable” (less than Electronic copy available at: https://ssrn.com/abstract=3954017 25 10% bedrock cover) or “eroded” (greater than 30% bedrock cover). Both AGB and AGB nutrient pools 26 increased with time since restoration at rates similar to coniferous plantations grown in areas 27 unimpacted by centuries of mining and smelting practices. Individual tree growth and nutrient 28 accumulation did not differ between stable and eroded sites; however, stable sites had a higher stem 29 density leading to overall higher AGB and AGB nutrient pools. Future N limitation of the regreening 30 forests does not appear to be a concern as aboveground N pools are six times larger than applied N, 31 indicating additional N is entering into the system whether through residual soil organic matter or the 32 establishment of N-fixing species. Conversely, aboveground P concentrations are decreasing with time 33 since tree planting and the 40-year-old study sites have aboveground P concentrations below values for 34 “healthy” trees. This study shows that the regreening efforts have led to a massive addition of 1, 144, 35 588 Mg of AGB (550, 547 Mg C) onto the landscape, and capable of sustaining healthy tree growth up to 36 40-years post restoration. However, as the regreening stands age, nutrient limitation may impact future 37 tree growth. Future studies should continue to investigate nutrient cycling within these remediated 38 forests particularly nutrients of concern such as P. 40 Keywords 41 forest restoration; degraded landscape; tree growth; forest nutrients; erosion 43 1. Introduction 44 Degraded landscapes with potential for forest restoration are estimated to total 1.7 – 1.8 billion 45 hectares of land and occur in almost every country in the world (Bastin et al. 2019). The 46 Intergovernmental Panel on Climate Change (IPCC) and the International Union for the Conservation of 47 Nature (IUCN) state that restoring forests on degraded landscapes will be essential to supporting global 48 carbon (C) mitigation measures and protecting biodiversity (IUCN, 2020; Rogelj et al. 2018). Most forest Electronic copy available at: https://ssrn.com/abstract=3954017 49 restoration projects do not fully recover pre-degradation ecosystem services and successful projects 50 require well-planned forest management (Holl & Brancalion, 2020). While large areas of degraded land 51 are ecologically capable of supporting forests, not all this land is available for forest restoration due to 52 complex land rights; an exception being mining and smelting degraded landscapes where forest 53 restoration is widely supported (Friedlingstein et al. 2019). 55 Soils on mining and smelting impacted landscapes are often heavily eroded, acidic, 56 contaminated with metals, and depleted of essential nutrients requiring external nutrient inputs to 57 support ecosystem function (Pietrzykowski & Daniels 2014). Applying crushed limestone and fertilizer 58 prior to tree planting has been found to increase soil pH, microbial biomass, and nutrients as well as 59 decrease the bioavailability of toxic metals (Kellaway et al. 2021; Narendrula-Kotha & Nkongolo, 2017). 60 Tree growth has also been found to increase after lime and fertilizer application compared with 61 unamended sites (Rumney et al. 2021). In forests, aboveground biomass (AGB) is a significant ecosystem 62 pool of essential nutrients, such as calcium (Ca), magnesium (Mg), nitrogen (N), phosphorous (P), 63 potassium (K), and C (Dielemen et al. 2020; Hooker & Compton 2003; Johnson & Todd 1998; Likens et al. 64 1994; Likens et al. 1998; Yanai 1992). Nutrients stored within AGB are not only essential for tree growth; 65 the cycling of AGB nutrients to other ecosystem pools provides necessary nutrients to all other forest 66 biota (Gordon & Jackson 2000; Gorgolewski et al. 2020; Neumann et al. 2018). 68 Aboveground biomass has traditionally been measured using field studies, which is regarded as 69 the most accurate way to estimate AGB at small scales (Lambert et al. 2005; Lu 2005). Landscape scale 70 AGB estimates are often conducted using a combination of field measurements and spatial 71 characteristics collected from remote sensing such as Landsat (Seidel et al. 2011; Lu et al. 2016). High 72 correlations between vegetation properties and spectral bands allow Landsat images to accurately Electronic copy available at: https://ssrn.com/abstract=3954017 73 estimate AGB across large areas (Lu 2005). Landsat images are free to access from the United States 74 Geological Survey’s (USGS) EarthExplorer Program anywhere across the globe, further increasing their 75 popularity (USGS 2020). Many AGB predictive models using Landsat images are developed on fully 76 forested sites (López-Serrano et al. 2020); or to assess the impacts of deforestation (Wu et al. 2016). 78 The City of Greater Sudbury, Ontario, Canada, and surrounding region has been exposed to 79 legacy effects of centuries of copper (Cu) and nickel (Ni) mining and smelting. Smelting emissions, along 80 with expansive clear-cutting led to 20,000 ha of barren land with highly eroded soils depleted of 81 essential nutrients (Gunn et al. 1995). Since 1978, billions of dollars have been invested by industry and 82 government to limit emissions and begin one of the world’s largest “regreening” projects. The 83 regreening landscape remains highly heterogenous including large areas of exposed bedrock with little 84 to no soil following massive erosion events. A recent study by Preston et al. (2020) measured AGB at 85 forested sites (n=10) across the regreening area; while Preston et al. (2020) did not incorporate historic 86 soil erosion into their site selection, they highlighted it’s potential to affect AGB growth. The objectives 87 of the current study are to: 1) assess how AGB and AGB nutrient pools change over time since 88 restoration at both highly eroded and stable sites using a space-for-time approach. 2) quantify AGB and 89 AGB nutrient pools arising from the City of Greater Sudbury regreening program. To account for the 90 effects of erosion on AGB growth and AGB nutrient accumulation we established two series of sites 91 ranging in stand ages. One series of sites was heavily influenced by erosion and largely covered by 92 exposed bedrock (>30% bedrock cover) (referred to as “eroded sites”); the other series was covered 93 with a layer of soil (>90% soil cover) (referred to as “stable sites”) (Figure 1). Aboveground biomass and 94 AGB nutrient pools were expected to increase with time since tree planting at both stable and eroded 95 sites, but pool sizes were expected be larger at stable sites. Field measurements from both eroded and Electronic copy available at: https://ssrn.com/abstract=3954017 96 stable sites along with spatial characteristics from Landsat images were then input into a random forest 97 (RF) model to estimate AGB and AGB nutrient pools across the City of Greater Sudbury regreening areas. 99 Figure 1. On the left is a picture of CON4 an eroded site with an average mineral soil depth of 4.9 cm. On 100 the right is a picture of CON1 a stable site with an average mineral soil depth of 32.6 cm. 102 2. Materials and Methods 103 2.1 Study area and regreening process 104 The City of Greater Sudbury (UTM zone 17; Easting: 501620, Northing: 5148797) is in 105 northeastern Ontario, Canada on the southern portion of the Precambrian Shield. Average monthly 106 temperature ranges from -13.0 ⁰C in January to 19.1 ⁰C in July, and average precipitation is 904 mm per 107 year (Environment Canada 2019). The City of Greater Sudbury lies in the vegetation transition zone 108 between the Great Lakes – St. Lawrence Forests regions of the south and the boreal forest to the north 109 (Rowe 1972). The topography of the region is undulating and dominated by rocky outcrops, soils are 110 typically shallow and sandy (Pearson & Pitbaldo 1995). Prior to regreening efforts, the soils within the 111 barrens of the City of Greater Sudbury were highly acidic, devoid of nutrients (particularly P, K, Ca, and 112 Mg) and contained high concentrations of bioavailable Ni and Cu (Winterhalder 1995). In order to 113 increase pH and soil nutrient concentrations over 3,400 ha of land were treated with 10 Mg/ha of 114 crushed dolomitic limestone (CaMg(CO ) ), 390 kg/ha of 6-24-24 N-P-K fertilizer. A grass-seed mixture, 3 2 Electronic copy available at: https://ssrn.com/abstract=3954017 115 that included N-fixing legumes, was applied at a rate of 40 kg/ha after liming and fertilizing. Across the 116 entire City of Greater Sudbury, 9.9 million trees have been planted, planted trees were predominantly 117 Jack pine (Pinus banksiana) and Red pine (Pinus resinosa), but also included white pine (Pinus strobus), 118 white spruce (Picea glauca) and green alder (Alnus viridis) (VETAC, 2020). 120 Figure 2. Map of the City of Greater Sudbury, Ontario, Canada including sixteen study sites, the outlines 121 for the barren and semi-barren regions, and the location of the three main smelters. 123 2.2 Site Selection 124 Sixteen sites varying in age were established within the barrens of the Coniston and Copper Cliff 125 smelters (Figure 2). The barrens were defined as areas that were devoid of plant life and had soil pH 126 below 4.3 before regreening efforts took place (Lautenbach et al. 1995). All sites were limed, fertilized, Electronic copy available at: https://ssrn.com/abstract=3954017 127 seeded, and planted with either red pine and/or jack pine. Planting dates ranged from 1979 to 2006, and 128 sites that only received planting at one time were specifically selected. Sites were split into two erosion 129 classes; highly eroded sites had greater than 30% bedrock exposure and typically had very little soil. 130 Sites with less than 10% bedrock exposure and deeper soils were considered “stable”. Six eroded sites 131 and ten stable sites were included in the study sites. Each site was 30 m x 30 m in size within a larger 132 forested area with similar characteristics. 134 2.3 Field sampling 135 Field sampling was conducted in June 2021. At each site, diameter at breast height (dbh) and 136 tree species were recorded for each tree. Trees were categorized as woody plants with a dbh larger than 137 2.5 cm. Soil depth was measured at nine points at each site using a soil auger. Forest floor depths were 138 taken at nine points by separating the forest floor from mineral soil using a knife. Elevation was 139 measured at each site using a Garmin eTrex Legend H GPS. Slope was measured at each site using a 140 Suunto PM-5/1520 clinometer. Tree cores were taken from three trees in each plot to verify age and 141 used for wood nutrient analysis. Two cores were taken from each sampled tree, one at breast height 142 and the other 20 cm below at a 90⁰ angle from the entrance point of the initial core. Cores were stored 143 in plastic straws with the ends taped. Bark samples were taken from these same trees at breast height 144 by scraping bark off the tree using a knife and composited from the three trees in a polyethylene bag. 145 Foliage samples were taken by hand from highest possible branch for 3 – 5 individual trees and 146 composited into paper bags. Current year needles were targeted for foliage sampling. Most sites were 147 monocultures of jack pine or red pine; cores, bark and foliage samples were only completed for the 148 dominant species at each site. On the occasion that the stands had a more heterogenous species 149 composition, samples were taken for any species that comprised greater than 33% of individuals per Electronic copy available at: https://ssrn.com/abstract=3954017 150 site. All vegetation (cores, bark, and foliage) samples were stored in a refrigerator (4 ⁰C) upon return 151 from the field. Vegetation samples were only collected for eleven of the sixteen study sites. 153 2.4 Chemical analysis 154 All vegetation samples were milled to a fine powder using a coffee grinder. Milled samples were 155 analyzed for Ca, Mg, K and P by digesting 0.2 g of sample in 2.5 mL of concentrated nitric acid (trace 156 grade 70% nitric acid (HNO )) at 100 ⁰C for eight hours followed by a digestion at room temperature for 157 eight hours. The digested material was filtered through a 0.45 µm pore filter paper and diluted using B- 158 pure until total volume reached 25 mL. The digest samples were then diluted 1:10 using B-pure and 159 analyzed using a Perkin Elmer Inductively Ion Coupled Optical Emission Spectrometer (7000 DV). Carbon 160 and N values were determined for the milled vegetation samples using an Elementar Macro CNS 161 analyzer. Quality control was confirmed by running all samples in triplicate and including experimental 162 blanks as well as a NIST-1515 apple leaf standard every 25 samples. 164 2.5 Aboveground biomass and nutrient pools 165 Aboveground biomass was calculated for each site using dbh-based allometric equations from 166 Lambert et al. (2005). Aboveground nutrient (Ca, Mg, N, P, K, and C) pools were determined using 167 measured nutrient concentrations for each biomass component (bark, foliage, wood + branches) and 168 the mass of each biomass component from the allometric equations. Tree core chemistry was used for 169 both wood and branches biomass components. If a site contained multiple species with greater than 170 33% individuals than chemistry for both red and jack pine were averaged. Aboveground biomass pools 171 were calculated for all sixteen study sites, aboveground nutrient pools were calculated for the eleven 172 sites where chemical analyses on vegetation samples were conducted. Aboveground biomass and Electronic copy available at: https://ssrn.com/abstract=3954017 173 nutrient accumulation rates were determined by dividing AGB or nutrient pool size by years since tree 174 planting. 176 2.6 Landsat image processing 177 The Landsat image was acquired from USGS Landsat 8 OLI program from July 9, 2020 (path: 19 178 and row: 28) level 1T (L1T), an image was selected that had less than 5% cloud cover. Landsat Level-1 179 products are geometrically corrected by orthorectification of ground control points and digital elevation 180 models (USGS 2019). The image was radiometrically and atmospherically corrected using the semi- 181 classification plugin based on the dark object subtraction (DOS1) technique (Chavez 1996) on Quantam 182 GIS (QGIS) Version 3.2.1 (QGIS.org 2020). 184 2.7 Random forest model development and application 185 A RF model was developed using Landsat 8 OLI bands 2 – 7, and ten vegetation indices (VIs) 186 (Table 1) at 35 sites within the City of Greater Sudbury barren areas where AGB was measured. Each 187 remote sensing variable included in model development has previously been shown to reflect various 188 components of AGB (Chenge & Osho, 2018; López-Serrano et al. 2020; Zhang et al. 2019). The 35 sites 189 used for model development included: the sixteen study sites, ten sites planted with red and/or jack 190 pine from Rumney et al. (2021), and nine sites that had an AGB of zero (three areas with bare exposed 191 bedrock, three roads, and three lakes). The study sites and zero AGB sites were 30 m by 30 m squares 192 while the Rumney et al. (2021) sites were circular plots with a diameter of 35.68 m. 194 To assess AGB and AGB nutrient pools for the regreening areas the RF model was applied to the 195 19,649 ha where red and/or jack pine had been planted across the barren areas of the City of Greater 196 Sudbury. Nutrient pools were determined for 7,216 ha of the barren areas where red and/or jack pine Electronic copy available at: https://ssrn.com/abstract=3954017 197 were planted and liming, and fertilizing took place. Magnesium, K, and C AGB concentrations were 198 determined by averaging the AGB nutrient concentrations from the study sites. Calcium, P, and N AGB 199 concentrations were determined using a linear curve based on the results from the study sites where 200 nutrient concentration varied with time since tree planting. Location and dosage rate of lime and 201 fertilizer applications were available from the City of Greater Sudbury regreening data (City of Greater 202 Sudbury 2020). 204 Table 1. Vegetation indices included in random forest to develop predictive model for aboveground 205 biomass using remotely sensed data. Landsat 8 OLI bands depicted using b. L is soil brightness correction 206 factor and its value is 0.5. Vegetation Index Equation Reference Brightness index (BI) Baig et al. (2014) 𝑏2 ∗ 0.3029 + 𝑏3 ∗ 0.2786 + 𝑏4 ∗ 0.4733 + 𝑏5 ∗ 0.5599 + 𝑏6 ∗ 0.508 + 𝑏7 ∗ 0.1872 Green normalized difference vegetation index (GNDVI) (𝑏5 ‒ 𝑏3)/(𝑏5 + 𝑏3) Gitelson et al. (1996) Greenness vegetation index (GVI) 𝑏2 ∗‒ 0.2941 + 𝑏3 ∗‒ 0.243 + 𝑏4 ∗‒ 0.5424 + 𝑏5 ∗ 0.7276 + 𝑏6 ∗ 0.0713 + 𝑏7 ∗‒ 0.1608 Baig et al. (2014) Land surface water index (LSWI) Xiao et al. (2004) (𝑏5 ‒ 𝑏6)/(𝑏5 + 𝑏6) Modified soil adjusted vegetation index (MSAVI) Qi et al. (1994) ((𝑏5 ‒ 𝑏4)/(𝑏5 + 𝑏4)) ∗ (1 + 𝐿) Normalized difference infrared index (NDII) (𝑏5 ‒ 𝑏6)/(𝑏5 + 𝑏6) Kimes et al. (1981) Normalized difference infrared index with SWIR2 (NDII7) (𝑏5 ‒ 𝑏7)/(𝑏5 + 𝑏7) Kimes et al. (1981) Normalized difference vegetation index (NDVI) (𝑏5 ‒ 𝑏4)/(𝑏5 + 𝑏4) USGS, (2016) Normalized difference water index (NDWI) Gao, (1996) (𝑏3 ‒ 𝑏6)/(𝑏3 + 𝑏6) Wetness index (WI) 𝑏2 ∗ 0.1511 + 𝑏3 ∗ 0.1973 + 𝑏4 ∗ 0.3283 + 𝑏5 ∗ 0.3407 + 𝑏6 ∗‒ 0.7117 + 𝑏7 ∗‒ 0.4559 Baig et al. (2014) 208 2.8 Statistical analysis 209 The relationship between AGB (at the hectare and stem scale), and AGB nutrient concentrations 210 with time since tree planting was assessed using regression coefficient (R ) values from linear regression. 211 Normality was tested for using the Lilliefors corrected Kolmogorov-Smirnov test, variables were 212 considered normally distributed if p was greater than 0.05. Differences between erosion classes in AGB 213 and nutrient accumulation rates were compared using an independent two-group Mann-Whitney- 214 Wilcoxon test; significant differences were determined if p was less than 0.05. The AGB predictive model 215 was developed using RF regression for seventeen remotely sensed predictor variables and AGB 216 measurements from 35 sites. Random forest was conducted through the Forest-based Classification and Electronic copy available at: https://ssrn.com/abstract=3954017 217 Regression function in ArcGIS Pro version 2.8.2. (ArcGIS Pro, 2021). The training of the RF model was 218 conducted on 90% of the available data with 10% being held back for validation. The selected RF model 219 included 1500 decision trees, 5 randomly selected variables, and 4 nodes per tree. The RF model was 220 evaluated on the training data set using residual mean squared error (RMSE) and regression coefficient 2 2 221 (R ) values and validated using the R value from the validation data set as well as out-of-bag RMSE. All 222 statistical analysis was done on R version 4.0.2 (RStudio Team 2020). All maps were created using 223 Quantam GIS (QGIS) Version 3.2.1 (QGIS.org 2020). 225 3. Results 226 Total stand age ranged from 14 to 41 years old with a median age of 27 years old (Table 2). 227 Median eroded site age was 27 years old while median stable site age was 25 years old. Fifty percent of 228 the total sites were dominated by jack pine, 25% by red pine and 25% were a mix of red and jack pine 229 (Table 2). Stem density ranged from 578 to 2,378 stems/ha with a median stem density of 1,660 230 stems/ha. Eroded sites had a median stem density of 1,144 stems/ha while stable sites had a stem 231 density of 1,668 stems/ha. Average tree dbh ranged from 5.1 cm to 16.0 cm with a median of 9.5 cm. 232 Both eroded and stable sites had a median tree dbh of 9.5 cm. Soil depth was a median of 25.06 cm 233 deep for stable sites and 6.00 cm for eroded sites. 235 Table 2. Site characteristics of sixteen study sites within the barren areas of the City of Greater Sudbury, 236 Ontario, Canada used to develop relationship between aboveground biomass and stand age. Tree Forest Mineral density floor soil Age Erosion Elevation Dominant (stems/ DBH depth depth Site Easting Northing (years) class (m) Slope (⁰) species ha) (cm) (cm) (cm) CC1 498390 5148618 24 Stable 309 4 Jack pine 1,822 14.3 4.58 15.06 CC2 498513 5148661 32 Eroded 300 15 Jack pine 1,866 8.1 1.07 3.56 CC3 498375 5150759 23 Stable 298 0 Red pine 1,644 10.9 1.32 25.87 CC10 495810 5143521 37 Stable 215 8 Red pine 1,655 14.7 5.00 23.33 Electronic copy available at: https://ssrn.com/abstract=3954017 CON1 515267 5148285 34 Stable 270 15 Jack pine 1,078 14.3 4.35 32.62 CON3 513804 5148482 40 Stable 211 8 Jack pine 1,744 8.1 1.78 14.31 CON4 513724 5148484 32 Eroded 258 13 Jack pine 1,666 8.8 0.88 4.94 CON5 512289 5149244 28 Stable 268 3 Mixed 1,666 7.4 4.25 22.94 CON6 512320 5149231 28 Eroded 281 6 Red pine 889 13.2 1.63 10.06 CON7 511116 5147545 21 Stable 241 0 Mixed 2,378 7.0 1.37 46.40 CON9 509020 5149553 15 Eroded 304 6 Jack pine 1,211 5.1 1.03 6.00 CON10 509140 5149646 15 Stable 290 14 Mixed 1,500 6.5 0.92 25.06 CON12 513979 5148306 42 Eroded 278 17 Jack pine 578 16.0 3.50 6.22 CON15 508894 5149251 24 Stable 215 14 Mixed 1,711 7.7 5.50 23.33 CON16 511895 5149249 34 Stable 278 2 Red pine 2,178 14.5 7.00 24.33 D1 509127 5145262 22 Eroded 295 2 Jack pine 1,078 10.2 1.33 4.87 Electronic copy available at: https://ssrn.com/abstract=3954017 238 239 Figure 3. Relationship between aboveground biomass (AGB) and time since planted for both highly 240 eroded sites with >30% rock cover and stable sites (soil cover >80%). Figure 3A depicts relationship at 241 the hectare scale; figure 3B is per individual stem. Dashed lines represent 95% confidence intervals. 242 Regression coefficient (R ) for relationship between AGB and age for both erosion classes is included. 244 Aboveground biomass (Mg/ha) had a significant positive linear relationship with time since tree 2 2 245 planting at the site scale for both the eroded (R =0.97, p<0.01) and stable sites (R =0.91, p<0.01) (Figure Electronic copy available at: https://ssrn.com/abstract=3954017 246 3A). Stable sites (2.78 ± 1.05 Mg/ha/year) had a significantly larger (p<0.01) AGB growth rate compared 247 with eroded sites (1.55 ± 0.38 Mg/ha/year). Aboveground biomass (kg/stem) also had a significant 248 positive linear relationship with time since tree planting at the stem scale for eroded (R =0.78) and 249 stable sites (R =0.82), however, AGB growth rates per stem were not significantly different (p=0.30) 250 between eroded (1.34 ± 0.71 kg/stem/year) and stable sites (1.59 ± 0.85 kg/stem/year) (Figure 3B). 252 Table 3. Aboveground biomass nutrient concentrations and accumulation rates (at both the hectare and 253 stem scale). Asterix denotes significant difference in nutrient concentration, or accumulation rate 254 between eroded and stable erosion classes (p<0.05). Nutrient concentrations Accumulation Rate Accumulation Rate Nutrient Erosion Class (g/kg) (kg/ha/year) (g/stem/year) Calcium Eroded 1.44 ± 0.51 1.80 ± 0.29 1.49 ± 0.34 Calcium Stable 1.10 ± 0.21 3.06 ± 0.76 1.61 ± 0.88 Magnesium Eroded 0.31 ± 0.08 0.48 ± 0.16 0.26 ± 0.20 Magnesium Stable 0.26 ± 0.05 0.74 ± 0.23 0.41 ± 0.19 Nitrogen Eroded 2.07 ± 0.35 2.77 ± 0.69* 2.34 ± 0.76 Nitrogen Stable 2.22 ± 0.43 5.18 ± 1.77* 2.90 ± 2.10 Phosphorous Eroded 0.14 ± 0.05 0.18 ± 0.04* 0.13 ± 0.07 Phosphorous Stable 0.15 ± 0.06 0.37 ± 0.12* 0.24 ± 0.07 Potassium Eroded 0.61 ± 0.17 0.75 ± 0.16* 0.67 ± 0.18 Potassium Stable 0.67 ± 0.15 1.51 ± 0.38* 0.90 ± 0.34 Carbon Eroded 483.53 ± 3.16 726.67 ± 199.51* 574.17 ± 188.70 Carbon Stable 485.62 ± 6.36 1279.70 ± 402.59* 669.08 ± 475.10 258 Aboveground nutrient concentrations were not significantly different between eroded or stable 259 sites (p>0.05) (Table 3). Accumulation rates of N (p<0.01), P (p<0.01), K (p<0.01), and C (p<0.05) within 260 AGB were significantly higher for stable versus eroded sites at the hectare scale. However, at the stem 261 scale there was no significant difference between nutrient accumulation rates and site erosion class 262 (p>0.05) (Table 3). There was a significant negative linear relationship between AGB concentration of Ca 2 2 2 263 (R =-0.41, p<0.05) (Figure 4A), N (R =0.-60, p<0.01) (Figure 4B), and P (R =-0.62, p<0.01) (Figure 4C) and Electronic copy available at: https://ssrn.com/abstract=3954017 264 time since tree planting and no relationship between concentration of Mg, K, or C in AGB and time since 265 tree planting (p>0.05) (Table 4). There was a significant positive linear relationship between time since 266 tree planting and AGB pools of all measured nutrients (p<0.05) (Table 4). Electronic copy available at: https://ssrn.com/abstract=3954017 267 268 Figure 4. Relationship between aboveground nutrient concentrations and time since tree planting for 269 calcium (Ca) (Figure 4A), nitrogen (N) (Figure 4B), and phosphorous (P) (Figure 4C). Dashed lines 270 represent 95% confidence intervals. Regression coefficient (R ) is included. Electronic copy available at: https://ssrn.com/abstract=3954017 2 272 Table 4. Regression coefficients (R ) for the relationships between aboveground biomass (AGB) nutrient 273 concentrations and time since tree planting, and the relationship between AGB nutrient pool sizes and 274 time since tree planting. Asterix denotes significant linear relationship (p<0.05). AGB nutrient AGB nutrient pool concentration (g/kg) size (g/stem) and and time since tree time since tree 2 2 Nutrient planting (R ) planting (R ) Calcium -0.41* 0.78* Magnesium -0.10 0.80* Nitrogen -0.62* 0.67* Phosphorous -0.60* 0.78* Potassium 0.24 0.82* Carbon -0.01 0.72* 278 Figure 5. Predicted aboveground biomass (AGB) values compared with measured AGB for 35 sites where 279 AGB was calculated. Black solid line represents 1:1 line. Solid green line represents regression line. 280 Dashed green lines represent 95% confidence intervals. Relationship between measured and predicted 281 AGB values assessed using root mean square error (RMSE) along with regression coefficient (R ). Electronic copy available at: https://ssrn.com/abstract=3954017 282 283 Aboveground biomass predictions from the RF model had a R of 0.91 and RMSE of 13.6 Mg/ha 284 when compared to AGB measurements from the training data set. Validation of the RF model revealed a 285 R of 0.91 when compared to AGB measurements of the validation data set and an out-of-bag RMSE of 286 31.35 Mg/ha. Total AGB predicted using the RF model for red and jack pine planted sites (19,649 ha) of 287 the City of Greater Sudbury is 1, 144, 588 Mg (550, 547 Mg C) (Figure 6). The mass of applied Ca and Mg 288 (in the form of crushed limestone) were 20 – 100 times greater than the mass of applied N, P, or K (in 289 the form of 6-24-24 fertilizer) for the 5,848 ha of limed, fertilized, and reforested land (Table 5). The 290 mass of N stored in AGB was over six times as high as the mass applied. The proportion of P and K stored 291 within AGB compared to applied nutrients was also substantial (Table 5). Electronic copy available at: https://ssrn.com/abstract=3954017 294 Figure 6. Aboveground biomass for red pine and jack pine planting polygons across the City of Greater 295 Sudbury, Ontario, Canada. 297 Table 5. Comparison of applied nutrients (applied as either lime or fertilizer) with nutrients stored in 298 aboveground biomass (AGB). Aboveground biomass nutrient pools are for all sites in the City of Greater 299 Sudbury that were limed, fertilized, and planted with Red and/or Jack pine (5,848 ha). Proportion of applied Nutrient concentration nutrients stored in Nutrients Applied (Mg) (g/kg) AGB (%) Calcium 12,707 1.25 ± 0.41 3.74 Magnesium 7707 0.28 ± 0.07 1.17 Nitrogen 126 2.16 ± 0.40 614.41 Phosphorous 239 0.14 ± 0.06 24.84 Potassium 454 0.64 ± 0.16 45.15 302 4. Discussion 303 Greater AGB pools at stable versus eroded sites was initially assumed to be due to deeper soils 304 leading to increased stem growth, as both jack pine and red pine have been found to be more 305 productive in deeper soils (Rudolph & Laidly 1990; Rudolph 1990). However, this was not the case as 306 individual stem growth did not differ between the two erosion classes and differences in hectare scale 307 AGB totals is driven primarily by stem density. Initial planting densities reported by the City of Greater 308 Sudbury (2020) were similar to measured stem density at the study sites, suggesting that the higher 309 stem density on stable sites is likely due to increased planting density as opposed to other factors such 310 as increased survivorship. Previous studies assessing C accumulation rates of jack and red pine 311 plantations in northern Ontario, Canada (Hunt et al. 2010), and Manitoba, Canada (Park, 2015), 312 converted AGB growth to C accumulation rates assuming C comprised 48% of AGB for Red pine and 50% 313 for Jack pine (Hunt et al. 2010; Park 2015). The individual stem growth rates from both Hunt et al. (2010) 314 (1.33 ± 0.33 kg/stem/year) and Park (2015) (0.92 ± 0.53 kg/stem/year) were very similar to values Electronic copy available at: https://ssrn.com/abstract=3954017 315 measured in the current study. Despite the history of highly acidified soils, high total metal 316 concentrations the trees in the regreening sites of the City of Greater Sudbury are growing at a similar 317 rate to those of jack and red pine plantations in relatively undisturbed areas regardless of erosion class. 319 Aboveground Ca, Mg, N, P, and K concentrations did not differ between eroded and stable sites 320 and median concentrations were within a standard deviation of healthy red pine nutrient 321 concentrations included in the United States Department of Agriculture Tree Chemistry Database (TCD) 322 (Pardo et al. 2005) (Table 6). Carbon concentration for wood at the study sites was 48.1 ± 0.6%, firmly in 323 between the Martin et al. (2018) values for boreal (46.8 ± 0.6%) and temperate conifers (50.1 ± 0.4%). 324 While the median AGB nutrient concentrations for the study sites are similar to the healthy values from 325 the TCD, concentrations of Ca, P, and N are decreasing with increasing time since tree planting. A 326 decrease in total AGB concentration of Ca, N, and P with increasing stand age has been documented 327 previously as these nutrients are stored in high concentrations in foliage which accumulates mass very 328 quickly in young trees as opposed to the slower growing, more nutrient poor wood (MacLean and Wein, 329 1977). The magnitude of the decrease for Ca and P at the study sites is cause for some concern as trees 330 at the two oldest study sites (CON1 and CON3) have bark and wood Ca and P concentrations well below 331 TCD values. Conversely, despite the decreasing concentration of N with time since tree planting, AGB N 332 concentration at CON1 and CON3 still exceed TCD values. 334 Table 6. Nutrient concentrations for each biomass component. Both the measured concentrations from 335 the study sites and the concentrations listed for Red pine from the Tree Chemistry Database (TCD) 336 (Pardo et al. 2005) are included. Bark Foliage Wood Study sites Study sites Study sites Nutrient (g/kg) TCD (g/kg) (g/kg) TCD (g/kg) (g/kg) TCD (g/kg) Electronic copy available at: https://ssrn.com/abstract=3954017 Ca 3.78 ± 2.17 7.65 ± 3.91 3.67 ± 1.81 4.17 ± 1.60 0.80 ± 0.15 1.09 ± 0.35 Mg 0.31 ± 0.14 0.46 ± 0.13 1.27 ± 0.62 0.86 ± 0.34 0.19 ± 0.05 0.19 ± 0.12 N 4.82 ± 1.28 3.10 ± 0.38 10.46 ± 1.16 11.49 ± 1.30 1.24 ± 0.40 0.82 ± 0.11 P 0.33 ± 0.20 0.44 ± 0.08 1.27 ± 0.33 1.28 ± 0.16 0.03 ± 0.03 0.08 ± 0.02 K 0.47 ± 0.47 0.88 ± 0.27 4.36 ± 1.59 3.67 ± 1.00 0.35 ± 0.12 0.24 ± 0.09 C 525.92 ± 26.15 Not available 492.50 ± 15.73 Not available 480.59 ± 5.90 Not available 338 Aboveground nutrient accumulation rates do not differ between erosion classes once stem 339 density is controlled for. Accumulation rates of Ca, Mg, N, P, K, and C for the study sites, increase over 340 time with a pattern similar to AGB growth and closely reflect values from coniferous plantations in 341 similar climates (Alban et al. 1978; Hunt et al., 2010; MacLean and Wein, 1977; Morrison, 1973; Park, 342 2015). Morrison (1973) assessed nutrient accumulation in a 30-year-old jack pine plantation in northern 343 Ontario, Canada at both high- and low-quality sites defined by biomass yield. The accumulation rates of 344 AGB Ca, Mg, P, and K at the City of Greater Sudbury regreening study sites closely reflected stands from 345 Morrison (1973) grown on poorer quality soils, while AGB N accumulation rates are more similar to the 346 higher quality sites. The similarity of the AGB growth rates, nutrient concentrations, and nutrient 347 accumulation rates to plantations unimpacted by mining and smelting practices suggest soil 348 amelioration has generally been a success in the City of Greater Sudbury barren areas. Soils in this area 349 were once inhospitable and now appear to support tree growth and nutrient accumulation similar to 350 poor quality but natural soils that are native to the southern portion of the Precambrian Shield. 352 A RF model incorporating ten VIs and Landsat 8 OLI bands 2 – 7 proved to be a successful 353 predictor of AGB for the study sites with higher R values and lower RMSE compared to studies using 354 similar methods (Gao et al. 2018; Jiang et al. 2021; López-Serrano et al. 2020; Zhang et al. 2019). The 355 stronger relationship and lower error for the model in the current study could be attributed to the 356 regreening forests being a low complexity stands, often a monoculture with limited understory; it has 357 been reported that models using VIs have decreased accuracy with high complexity forests (Lu et al. Electronic copy available at: https://ssrn.com/abstract=3954017 358 2016). An additional reason for the low RMSE of AGB estimates at the study sites, is the low AGB of 359 these forests, Landsat regression-based models have been found to become saturated at AGB values of 360 150 Mg/ha (Lu 2005). 362 The total AGB for red and jack pine stands across 19,649 ha within the City of Greater Sudbury is 363 1, 144, 588 Mg. Using an average of the eroded and stable site growth-curves the average AGB of 58.25 364 Mg/ha implies that the average site is 27.3 years old. As the remediated stands continue to age, we 365 expect AGB C to increase as jack pine accumulate biomass up to 100 years old and Red pines can 366 accumulate biomass until up to 300 years old (Rudolph & Laidly 1990; Rudolph 1990). The total AGB C 367 pool of 550, 547 Mg C is the equivalent to C emissions from 319, 158 cars over one year (3.67% of the 368 registered vehicles in the Province of Ontario) (NRCAN 2014). Applying the federal governments price 369 for C at $50 CAD per Mg, the AGB C pools of the regreening sites of the City of Greater Sudbury is equal 370 to $27.53 million CAD since the restoration project began in 1978. This total does not include any other 371 C pools such as mineral soil, forest floor, shrub, and herb pools. Preston et al. (2020) found that while 372 forest floor C increased with stand age, the mineral soil C pool’s relationship with stand age was unclear 373 and further work on the City of Greater Sudbury regreening sites suggests that soil C is decreasing with 374 stand age (Levasseur et al. unpublished). The current study’s AGB C projections only include sites 375 planted with jack and red pine, while Jack and Red pine are the most commonly planted tree species, 376 there are regions planted with other conifers and deciduous species that are not included in this total 377 (VETAC, 2020). 379 A concern for long-term forest growth on degraded landscapes is lack of nutrients. To overcome 380 the low nutrient concentrations of the City of Greater Sudbury’s barren soils, the regreening program 381 applied massive quantities of nutrients (particularly Ca and Mg) to the landscape (Winterhalder 1996). Electronic copy available at: https://ssrn.com/abstract=3954017 382 Only 4% of applied Ca and 1% of applied Mg were stored within AGB, this is most likely due to large- 383 scale erosion removing Ca and Mg from the landscape. A recent study conducted on the same sites as 384 the current study found that large quantities of applied Ca and Mg were unaccounted for in nutrient 385 budgets (Kellaway et al. 2021). High percentages of P, and K were accumulated within AGB which could 386 also be an indication of nutrient limitation. Phosphorous has often been reported to be the primary 387 limiting nutrient within the barrens of the City of Greater Sudbury (Beckett & Negusanti 1990; 388 Lautenbach et al. 1995; Munford et al. 2021). The limited soil organic matter, slow decomposition rates 389 caused a limited source of available P and the low soil pH caused large proportions of available P to be 390 bound to inorganic complexes (Marschner 1991). Previous studies on the City of Greater Sudbury 391 barrens found that when liming and fertilizer has been applied, vegetation rapidly accumulates P in 392 biomass (Winterhalder 1996). Little work has been done on K limitation in Sudbury; however, liming and 393 fertilizing has been found to either not impact or decrease bioavailable K in soils within the City of 394 Greater Sudbury (Nkongolo et al. 2013; SARA Group 2009). Despite the limited bioavailability of K in 395 soils, Kellaway et al. (2021) found that the equivalent of 95% of applied K at the study sites was stored in 396 biomass. 398 Nitrogen pools stored within AGB were found to be over six times as high as the amount applied 399 in fertilizer, implying trees were accumulating N from additional sources outside of fertilizer. The SARA 400 Group (2009) found that total N concentrations in forest soils of the City of Greater Sudbury barrens 401 were lower than those of reference forests, however, residual organic matter in the forest soils of the 402 barrens has been proposed to be a sufficiently large pool to sustain vegetation communities 403 (Winterhalder 1995). The pre-restoration forest soil N pools may make up the difference between N 404 applied as fertilizer and N stored in AGB. Additionally, N-fixing species such as Lotus corniculatus (Bird’s- 405 foot trefoil), Trifolium hybridum (Alsike clover) were planted widely during restoration, and the native Electronic copy available at: https://ssrn.com/abstract=3954017 406 Comptonia peregrina (Sweet fern), and Pleurozium sp. (Feather moss) have become common in the 407 recovering jack pine stands (Lautenbach et al. 1995). Limited recent work has been done on changes in 408 forest soil N pools within the City of Greater Sudbury, however, a study by Meyer-Jacob et al. (2020) 409 found that organic N has been increasing in the region’s lakes from 1984 – 2020. 411 While trees within the remediated sites appear to be accumulating N from additional sources 412 and have AGB N concentrations higher than healthy conifers this is not apparent for Ca, P, and K. 413 Calcium and P are of particular concern as both nutrients have declining concentrations with time since 414 tree planting and have concentrations lower than healthy conifers in the oldest study sites. As the 415 average age for the regreening sites is 27.3 years, existing Ca, P, and K pools, provided from fertilizer and 416 limestone may be depleted as these stands continue to age. An additional concern for future forest 417 growth is total metal concentrations in the regreening area, such as Ni and Cu, remain high in mineral 418 soils; however, bioavailable concentrations of these metals are very low (reported between 0.1 – 1% of 419 total metal concentrations) (Kellaway et al. 2021; Nkongolo et al. 2013). While the high total metal 420 concentrations do not appear to be affecting tree growth up to this point in the regreening program, 421 Kellaway et al. (2021) found that metal partitioning at the study sites was driven more by pH versus OM, 422 indicating that as effects of liming on pH diminish, metals may become more bioavailable potentially 423 impacting tree growth. 425 5. Conclusions 426 The City of Greater Sudbury regreening efforts are regarded as one of the world’s largest 427 restoration projects and have led to over one million tonnes of AGB (over 500, 000 tonnes of C) 428 establishing in the 19, 649 ha of regreening area over 40 years. Within the regreening area, individual 429 trees are growing and accumulating aboveground nutrients at rates similar to jack and red pine Electronic copy available at: https://ssrn.com/abstract=3954017 430 plantations grown in areas unimpacted by mining and smelting. Aboveground nutrient concentrations 431 exceeding TCD values (particularly N) similarly suggest that up to 40-years post restoration the 432 amendments applied to the City of Greater Sudbury regreening areas appear to be sufficient to support 433 healthy tree growth. However, some cause for concern going forward is aboveground P concentrations 434 are decreasing as trees age and 40-year old study sites already have aboveground P concentrations 435 below “healthy” tree values. Phosphorous limitation has been previously documented in the regreening 436 areas and future studies should continue to investigate potential future nutrient limitation. 438 Acknowledgements 439 This study is part of the Landscape Carbon Accumulation through Reductions in Emissions (L- 440 CARE) project and is funded by Ontario Centers for Excellence, Natural Sciences and Engineering 441 Research Council of Canada, Vale Canada Limited, Sudbury Integrated Nickel Operations and the City of 442 Greater Sudbury. The authors would like to thank the countless volunteers responsible for tree planting, 443 liming, and fertilizing the City of Greater Sudbury barrens. The authors would also like to thank the rest 444 of the L-CARE team as well as Kevin Adkinson, Kaylen Foley, Spencer Gilbert-Parkes, William Humphrey, 445 Edward Kellaway, Kimber Munford, and Jodi Newman for technical support throughout this study. The 446 authors declare that they have no conflict of interests. Electronic copy available at: https://ssrn.com/abstract=3954017 447 Literature cited 448 Alban DH, Perala DA, and Schlaegel BE (1978) Biomass and nutrient distribution in aspen, pine and 449 spruce stands on the same soil type in Minnesota. Canadian Journal of Forest Research 8: 290 – 450 299 451 ArcGIS Pro (2021) ESRI https://www.esri.com/en-us/arcgis/products/arcgis-pro/overview 452 Bastin JF, Finegold Y, Garcia C, Mollicone D, Rezende M, Routh D, Crowther, T. W. et al (2019). The 453 global tree restoration potential. Science 365: 76 – 79 454 Baig MHA, Zhang L, Shuai T, Tong Q (2014) Derivation of a tasselled cap transformation based on 455 Landsat 8 at-satellite reflectance. Remote Sensing Letters 5: 423 – 431 456 Beckett PJ, Negusanti J (1990) Using land reclamation practices to improve tree condition in the Sudbury 457 smelting area, Ontario, Canada. Proceedings of the 1990 Mining and Reclamation Conference 458 and Exhibition West Virginia University, Morgantown, West Virginia, USA 459 Chavez PS (1996) Image-based atmospheric corrections – revisited and improved. Photogrammetric 460 Engineering & Remote Sensing 62: 1025 – 1036 461 Chenge IB, Osho JSA (2018) Mapping tree aboveground biomass and carbon in Omo Forest Reserve 462 Nigeria using Landsat 8 OLI data. Southern Forests 80: 341 – 350 463 City of Greater Sudbury (2020) Regreening program https://www.greatersudbury.ca/live/environment- 464 and-sustainability1/regreening-program/ 465 Dieleman CM, Rogers BM, Potter S, Veraverbeke S, Johnstone JF, Laflamme J, Solvik K, Walker XJ, Mack 466 MC, Turetsky MR (2020) Wildfire combustion and carbon stocks in the southern Canadian boreal 467 forest: Implications for a warming world. Global Change Biology 26: 6062 – 6079 468 Environment Canada (2019) Canadian climate normals 1981 – 2010 station data 469 https://climate.weather.gc.ca/climate_normals/results_1981_2010_e.html Electronic copy available at: https://ssrn.com/abstract=3954017 470 Friedlingstein P, Allen M, Canadell JG, Peters GP, Seneviratne SI (2019) Comment on “The global tree 471 restoration potential”. Science 366: 1 – 2 472 Gao BC (1996) NDWI—A normalized difference water index for remote sensing of vegetation liquid 473 water from space. Remote sensing of environment 58: 257 – 266 474 Gao Y, Lu D, Li G, Wang G, Chen Q, Liu L, Li D (2018) Comparative analysis of modeling algorithms for 475 forest aboveground biomass estimates in a subtropical region. Remote Sensing 10: 627 476 Gitelson AA, Kaufman YJ, Merzlyak MN (1996) Use of a green channel in remote sensing of global 477 vegetation from EOS-MODIS. Remote sensing of Environment 58: 289 – 298 478 Gordon WS, Jackson RB (2000) Nutrient concentrations in fine roots. Ecology 81: 275 – 280 479 Gorgolewski A, Rudz P, Jones T, Basiliko N, Caspersen J (2020) Assessing coarse woody debris nutrient 480 dynamics in managed northern hardwood forests using a matrix transition model. Ecosystems 481 23: 541 – 554 482 Gunn J, Keller, W Negusanti J, Potvin R, Beckett P, Winterhalder K (1995) Ecosystem recovery after 483 emission reductions: Sudbury, Canada. Water, Air, & Soil Pollution 85: 1783 – 1788 484 Holl KD, Brancalion PHS (2020) Tree planting is not a simple solution. Science 368: 580 – 581 485 Hooker TB, Compton JE (2003) Forest ecosystem carbon and nitrogen accumulation during the first 486 century after agricultural abandonment. Ecological Applications 13: 299 – 313 487 Hunt SL, Gordon AM, Morris DM (2010) Carbon stocks in managed conifer forests in northern Ontario, 488 Canada. Silva Fennica 44: 563 – 582 489 International Union for the Conservation of Nature (IUCN) (2020) IUCN Annual 2019 Report. IUCN, 490 Gland, Switzerland. 491 Jiang F, Kutia M, Ma K, Chen S, Long J, and Sun H (2021). Estimating the aboveground biomass of 492 coniferous forest in Northeast China using spectral variables, land surface temperature and soil 493 moisture. Science of The Total Environment 785: 147335. Electronic copy available at: https://ssrn.com/abstract=3954017 494 Johnson DW, Todd DE (1998) Harvesting effects on long-term changes in nutrient pools of mixed oak 495 forest. Soil Science Society of America Journal 62: 1725 – 1735 496 Kellaway EJ, Eimers MC, and Watmough SA (2021). Liming legacy effects associated with the world's 497 largest soil liming and regreening program in Sudbury, Ontario, Canada. Science of The Total 498 Environment: 150321. 499 Kimes DS, Markham BL, Tucker CJ, & McMurtrey III JE (1981) Temporal relationships between spectral 500 response and agronomic variables of a corn canopy. Remote Sensing of Environment 11: 401 – 501 411 502 Lambert M-C, Ung C-H, Raulier F, (2005) Canadian national tree aboveground biomass equations 503 Canadian Journal of Forest Research 35: 1996 – 2018 504 Lautenbach WE, Miller J, Beckett PJ, Negusanti JJ, Winterhalder K (1995) Municipal land restoration 505 program: the regreening process. Pages 109 – 122 In: Gunn J (ed) Restoration and recovery of an 506 industrial region. Springer, New York, USA 507 Likens GE, Driscoll CT, Busco DC, Siccama TG, Johnson CE, Lovett GM, Fahey TJ, Reiners WA, Ryan DF, 508 Martin CW, Bailey SW (1998) The biogeochemistry of calcium at Hubbard Brook. 509 Biogeochemistry 41: 89 – 173 510 Likens GE, Driscoll CT, Busco DC, Siccama TG, Johnson CE, Lovett GM, Ryan DF, Fahey TJ, Reiners WA 511 (1994) The biogeochemistry of potassium at Hubbard Brook. Biogeochemistry 25: 61 – 125 512 López-Serrano PM, Cárdenas Domínguez JL, Corral-Rivas JJ, Jiménez E, López-Sánchez CA, Vega-Nieva DJ 513 (2020) Modeling of Aboveground Biomass with Landsat 8 OLI and Machine Learning in 514 Temperate Forests. Forests 11: 11 515 Lu D (2005) Aboveground biomass estimation using Landsat TM data in the Brazilian Amazon. 516 International Journal of Remote Sensing 26: 2509 – 2525 Electronic copy available at: https://ssrn.com/abstract=3954017 517 Lu D, Chen Q, Wang G, Liu L, Li G, Moran E (2016) A survey of remote sensing-based aboveground 518 biomass estimation methods in forest ecosystems. International Journal of Digital Earth 9: 63 – 519 105 520 MacLean DA, Wein RW (1977) Nutrient accumulation for postfire jack pine and hardwood succession 521 patterns in New Brunswick. Canadian Journal of Forest Research 7: 562 – 578 522 Marschner H (1991) Mechanisms of plant adaptation to acid soils. Plant and Soil 134: 1 – 20 523 Martin AR, Doraisami M, Thomas SC (2018) Global patterns in wood carbon concentration across the 524 world’s trees and forests. Nature Geoscience 11: 915 – 920 525 Meyer-Jacob C, Labaj AL, Paterson AM, Edwards BA, Keller W, Cummings BF, Smol JP (2020) Re- 526 browning of Sudbury (Ontario, Canada) lakes now approaches pre-acid deposition lake-water 527 dissolved organic carbon levels. Science of the Total Environment: 138347 528 Morrison IK (1973) Distribution of elements in aerial components of several natural jack pine stands in 529 northern Ontario. Canadian Journal of Forest Research 3: 170 – 179 530 Munford KE, Casamatta M, Basiliko N, Glasauer S, Mykytczuk NCS, Watmough SA (2021) Paper birch 531 (Betula papyrifera) nutrient resorption rates on nutrient-poor metal-contaminated soils and 532 mine tailings. Water, Air, & Soil Pollution 232: 33 533 Narendrula-Kotha R, and Nkongolo KK (2017). Microbial response to soil liming of damaged ecosystems 534 revealed by pyrosequencing and phospholipid fatty acid analyses. PloS one 12: e0168497. 535 Natural Resources Canada (NRCAN) (2014) Learn the facts: fuel consumption and CO 536 https://www.nrcan.gc.ca/sites/www.nrcan.gc.ca/files/oee/pdf/transportation/fuel-efficient- 537 technologies/autosmart_factsheet_6_e.pdf 538 Neumann M, Ukonmaanaho L, Johnson J, Benham S, Vesterdal L, Novotný R, Verstraeten A, Lundin L, 539 Thimonier A, Michopoulos P, Hasenauer H (2018) Quantifying carbon and nutrient input from Electronic copy available at: https://ssrn.com/abstract=3954017 540 litterfall in European forests using field observations and modeling. Global Biogeochemical 541 Cycles. 32: 784 – 798 542 Nkongolo KK, Spiers G, Beckett P, Narendrula R, Theriault G, Tran A, Kalubi KN (2013) Long-term effects 543 of liming on soil chemistry in stable and eroded upland areas in a mining region. Water, Air, & 544 Soil Pollution 224: 1618 545 Pardo LH, Duarte N, Miller EK, Robin-Abbot M (2005) Tree chemistry database (Version 1). USDA Forest 546 Service, Northeastern Research Station 547 Park A (2015) Carbon storage and stand conversion in a pine-dominated boreal forest landscape. Forest 548 Ecology and Management 340: 70 – 81 549 Pearson DAB, Pitblado JR (1995) Geological and geographical setting. Pages 5 – 15 In: Gunn J (ed) 550 Restoration and recovery of an industrial region. Springer, New York, USA 551 Pietrzykowski M, Daniels WL (2014) Estimation of carbon sequestration by pine (Pinus sylvestris L) 552 ecosystems developed on reforested post-mining sites in Poland on differing mine soil 553 substrates. Ecological Engineering 73: 209 – 218 554 Preston MD, Brummell ME, Smenderovac E, Rantala-Sykes B, Rumney RHM, Sherman G, Basiliko N, 555 Beckett P, Hebert M (2020) Tree restoration and ecosystem carbon storage in an acid and metal 556 impacted landscape: Chronosequence and resampling approaches. Forest Ecology and 557 Management 463: 118012 558 QGIS.org (2020) QGIS Geographic Information System http://www.qgis.org 559 Qi J, Chehbouni A, Huete AR, Kerr YH, & Sorooshian S (1994) A modified soil adjusted vegetation 560 index. Remote sensing of environment 48: 119 – 126 561 Rogelj J, Shindell D, Jiang K, Fifita S, Forster P, Ginzburg V, Handa C, Kheshgi H, Kobayashi S, Kriegler E, 562 Mundaca L, Séférian R, Vilariño MV (2018) Chapter 2: Mitigation pathways compatible with 563 1.5°C in the context of sustainable development. In: Masson-Delmotte V, Zhai P, Pörtner, HO, Electronic copy available at: https://ssrn.com/abstract=3954017 564 Roberts D, Skea J, Shukla PR, Pirani A, Moufouma-Okia W, Péan C, Pidcock R, Connors S, 565 Matthews JBR, Chen Y, Zhou X, Gomis MI, Lonnoy E, Maycock T, Tignor M, Waterfield T (eds) 566 Global Warming of 1.5°C. An IPCC Special Report on the impacts of global warming of 1.5°C 567 above pre-industrial levels and related global greenhouse gas emission pathways, in the context 568 of strengthening the global response to the threat of climate change, sustainable development, 569 and efforts to eradicate poverty. Intergovernmental Panel on Climate Change. 570 Rowe JS (1972) Forest regions of Canada. Canadian Forest Service, Department of Environment, Ottawa, 571 Canada 572 RStudio Team (2020) RStudio: Integrated Development for R http://www.rstudio.com/ 573 Rudolf PO (1990) Pinus resinosa Ait Red pine. Silvics of North America 1: 442 – 455 574 Rudolph TD, Laidly, PR (1990) Pinus banksiana Lamb Jack pine. Silvics of North America 1: 280 – 293 575 Rumney RHM, Preston MD, Jones T, Basiliko N, Gunn J (2021) Soil amendment improves carbon 576 sequestration by trees on severely damaged acid and metal landscape, but total storage remains 577 low. Forest Ecology and Management 483: 118896. 578 SARA Group (2009) Summary of Volume III: Ecological risk assessment. Sudbury Soils Study. SARA Group, 579 Guelph, Canada 580 Seidel D, Fleck S, Leuschner C, Hammett T (2011) Review of ground-based methods to measure the 581 distribution of biomass in forest canopies. Annals of Forest Science 68: 225 – 244 582 US Geological Survey (USGS) (2019) Landsat levels of processing details https://www.usgs.gov/core- 583 science-systems/nli/landsat/landsat-levels-processing 584 US Geological Survey (USGS) (2016) Landsat surface reflectance-derived spectral indices 585 https://www.usgs.gov/land-resources/nli/landsat/landsat-surface-reflectance-derived-spectral- 586 indices 587 US Geological Survey (USGS) (2020) EarthExplorer https://earthexplorer.usgs.gov/ Electronic copy available at: https://ssrn.com/abstract=3954017 588 Vegetation Enhancement Technical Advisory Committee (VETAC) (2020) 2020 Annual Report: 589 Regreening Program. VETAC, Sudbury, Canada 590 Winterhalder K (1995) Dynamics of plant communities and soils in revegetated ecosystems: a Sudbury 591 case study. Pages 173 – 182 In: Gunn J (ed) Restoration and recovery of an industrial region. 592 Springer, New York, USA 593 Winterhalder K (1996) Environmental degradation and rehabilitation of the landscape around Sudbury, a 594 major mining and smelting area. Environmental Reviews 4: 185 – 224 595 Wu C, Shen H, Wang K, Shen A, Deng J, Gan M (2016) Landsat imagery-based aboveground biomass 596 estimation and change investigation related to human activities. Sustainability 8: 159 597 Xiao X, Hollinger D, Aber J, Goltz M, Davidson EA, Zhang Q, Moore B (2004) Satellite-based modelling of 598 gross primary production in an evergreen needleleaf forest. Remote Sensing of the Environment 599 89: 519 – 534 600 Yanai RD (1992) Phosphorous budget of a 70-year-old northern hardwood forest. Biogeochemistry 17: 1 601 – 22 602 Zhang R, Zhuo X, Ouyang Z, Avitabile V, Qi J, Chen J, Giannico V (2019) Estimating aboveground biomass 603 in subtropical forests of China by integrating multisource remote sensing and ground data. 604 Remote Sensing of the Environment 232: 111341 Electronic copy available at: https://ssrn.com/abstract=3954017 612 Electronic copy available at: https://ssrn.com/abstract=3954017 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png AgricSciRN Subject Matter Journals SSRN

World's Largest Regreening Project Promotes Healthy Tree Growth and Nutrient Accumulation Up to 40-Years Post Restoration

Download PDF
33 pages

Loading next page...
 
/lp/ssrn/world-s-largest-regreening-project-promotes-healthy-tree-growth-and-3qLCxx0CDC

References

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

Copyright
Copyright © 2021 Elsevier Inc.
eISSN
ARCH-4001
Publisher site
See Article on Publisher Site

Abstract

17 Mining and smelting degraded landscapes are characterised by heavily eroded, acidic soils that are 18 contaminated with toxic metals and depleted of essential nutrients. Restoring forests on these 19 landscapes has been highlighted to support carbon (C) mitigation measures and protect biodiversity. 20 Understanding how tree growth and aboveground nutrient accumulation changes following restoration 21 will be essential to planning future forest restoration projects. In this study, we assessed aboveground 22 biomass (AGB) and aboveground nutrient (calcium (Ca), magnesium (Mg), nitrogen (N), phosphorous (P), 23 potassium (K), and C) pools across as series of sites ranging in age from 15 to 42 years old; to determine 24 the effects of erosion on AGB and AGB nutrient pools each site was categorized as “stable” (less than Electronic copy available at: https://ssrn.com/abstract=3954017 25 10% bedrock cover) or “eroded” (greater than 30% bedrock cover). Both AGB and AGB nutrient pools 26 increased with time since restoration at rates similar to coniferous plantations grown in areas 27 unimpacted by centuries of mining and smelting practices. Individual tree growth and nutrient 28 accumulation did not differ between stable and eroded sites; however, stable sites had a higher stem 29 density leading to overall higher AGB and AGB nutrient pools. Future N limitation of the regreening 30 forests does not appear to be a concern as aboveground N pools are six times larger than applied N, 31 indicating additional N is entering into the system whether through residual soil organic matter or the 32 establishment of N-fixing species. Conversely, aboveground P concentrations are decreasing with time 33 since tree planting and the 40-year-old study sites have aboveground P concentrations below values for 34 “healthy” trees. This study shows that the regreening efforts have led to a massive addition of 1, 144, 35 588 Mg of AGB (550, 547 Mg C) onto the landscape, and capable of sustaining healthy tree growth up to 36 40-years post restoration. However, as the regreening stands age, nutrient limitation may impact future 37 tree growth. Future studies should continue to investigate nutrient cycling within these remediated 38 forests particularly nutrients of concern such as P. 40 Keywords 41 forest restoration; degraded landscape; tree growth; forest nutrients; erosion 43 1. Introduction 44 Degraded landscapes with potential for forest restoration are estimated to total 1.7 – 1.8 billion 45 hectares of land and occur in almost every country in the world (Bastin et al. 2019). The 46 Intergovernmental Panel on Climate Change (IPCC) and the International Union for the Conservation of 47 Nature (IUCN) state that restoring forests on degraded landscapes will be essential to supporting global 48 carbon (C) mitigation measures and protecting biodiversity (IUCN, 2020; Rogelj et al. 2018). Most forest Electronic copy available at: https://ssrn.com/abstract=3954017 49 restoration projects do not fully recover pre-degradation ecosystem services and successful projects 50 require well-planned forest management (Holl & Brancalion, 2020). While large areas of degraded land 51 are ecologically capable of supporting forests, not all this land is available for forest restoration due to 52 complex land rights; an exception being mining and smelting degraded landscapes where forest 53 restoration is widely supported (Friedlingstein et al. 2019). 55 Soils on mining and smelting impacted landscapes are often heavily eroded, acidic, 56 contaminated with metals, and depleted of essential nutrients requiring external nutrient inputs to 57 support ecosystem function (Pietrzykowski & Daniels 2014). Applying crushed limestone and fertilizer 58 prior to tree planting has been found to increase soil pH, microbial biomass, and nutrients as well as 59 decrease the bioavailability of toxic metals (Kellaway et al. 2021; Narendrula-Kotha & Nkongolo, 2017). 60 Tree growth has also been found to increase after lime and fertilizer application compared with 61 unamended sites (Rumney et al. 2021). In forests, aboveground biomass (AGB) is a significant ecosystem 62 pool of essential nutrients, such as calcium (Ca), magnesium (Mg), nitrogen (N), phosphorous (P), 63 potassium (K), and C (Dielemen et al. 2020; Hooker & Compton 2003; Johnson & Todd 1998; Likens et al. 64 1994; Likens et al. 1998; Yanai 1992). Nutrients stored within AGB are not only essential for tree growth; 65 the cycling of AGB nutrients to other ecosystem pools provides necessary nutrients to all other forest 66 biota (Gordon & Jackson 2000; Gorgolewski et al. 2020; Neumann et al. 2018). 68 Aboveground biomass has traditionally been measured using field studies, which is regarded as 69 the most accurate way to estimate AGB at small scales (Lambert et al. 2005; Lu 2005). Landscape scale 70 AGB estimates are often conducted using a combination of field measurements and spatial 71 characteristics collected from remote sensing such as Landsat (Seidel et al. 2011; Lu et al. 2016). High 72 correlations between vegetation properties and spectral bands allow Landsat images to accurately Electronic copy available at: https://ssrn.com/abstract=3954017 73 estimate AGB across large areas (Lu 2005). Landsat images are free to access from the United States 74 Geological Survey’s (USGS) EarthExplorer Program anywhere across the globe, further increasing their 75 popularity (USGS 2020). Many AGB predictive models using Landsat images are developed on fully 76 forested sites (López-Serrano et al. 2020); or to assess the impacts of deforestation (Wu et al. 2016). 78 The City of Greater Sudbury, Ontario, Canada, and surrounding region has been exposed to 79 legacy effects of centuries of copper (Cu) and nickel (Ni) mining and smelting. Smelting emissions, along 80 with expansive clear-cutting led to 20,000 ha of barren land with highly eroded soils depleted of 81 essential nutrients (Gunn et al. 1995). Since 1978, billions of dollars have been invested by industry and 82 government to limit emissions and begin one of the world’s largest “regreening” projects. The 83 regreening landscape remains highly heterogenous including large areas of exposed bedrock with little 84 to no soil following massive erosion events. A recent study by Preston et al. (2020) measured AGB at 85 forested sites (n=10) across the regreening area; while Preston et al. (2020) did not incorporate historic 86 soil erosion into their site selection, they highlighted it’s potential to affect AGB growth. The objectives 87 of the current study are to: 1) assess how AGB and AGB nutrient pools change over time since 88 restoration at both highly eroded and stable sites using a space-for-time approach. 2) quantify AGB and 89 AGB nutrient pools arising from the City of Greater Sudbury regreening program. To account for the 90 effects of erosion on AGB growth and AGB nutrient accumulation we established two series of sites 91 ranging in stand ages. One series of sites was heavily influenced by erosion and largely covered by 92 exposed bedrock (>30% bedrock cover) (referred to as “eroded sites”); the other series was covered 93 with a layer of soil (>90% soil cover) (referred to as “stable sites”) (Figure 1). Aboveground biomass and 94 AGB nutrient pools were expected to increase with time since tree planting at both stable and eroded 95 sites, but pool sizes were expected be larger at stable sites. Field measurements from both eroded and Electronic copy available at: https://ssrn.com/abstract=3954017 96 stable sites along with spatial characteristics from Landsat images were then input into a random forest 97 (RF) model to estimate AGB and AGB nutrient pools across the City of Greater Sudbury regreening areas. 99 Figure 1. On the left is a picture of CON4 an eroded site with an average mineral soil depth of 4.9 cm. On 100 the right is a picture of CON1 a stable site with an average mineral soil depth of 32.6 cm. 102 2. Materials and Methods 103 2.1 Study area and regreening process 104 The City of Greater Sudbury (UTM zone 17; Easting: 501620, Northing: 5148797) is in 105 northeastern Ontario, Canada on the southern portion of the Precambrian Shield. Average monthly 106 temperature ranges from -13.0 ⁰C in January to 19.1 ⁰C in July, and average precipitation is 904 mm per 107 year (Environment Canada 2019). The City of Greater Sudbury lies in the vegetation transition zone 108 between the Great Lakes – St. Lawrence Forests regions of the south and the boreal forest to the north 109 (Rowe 1972). The topography of the region is undulating and dominated by rocky outcrops, soils are 110 typically shallow and sandy (Pearson & Pitbaldo 1995). Prior to regreening efforts, the soils within the 111 barrens of the City of Greater Sudbury were highly acidic, devoid of nutrients (particularly P, K, Ca, and 112 Mg) and contained high concentrations of bioavailable Ni and Cu (Winterhalder 1995). In order to 113 increase pH and soil nutrient concentrations over 3,400 ha of land were treated with 10 Mg/ha of 114 crushed dolomitic limestone (CaMg(CO ) ), 390 kg/ha of 6-24-24 N-P-K fertilizer. A grass-seed mixture, 3 2 Electronic copy available at: https://ssrn.com/abstract=3954017 115 that included N-fixing legumes, was applied at a rate of 40 kg/ha after liming and fertilizing. Across the 116 entire City of Greater Sudbury, 9.9 million trees have been planted, planted trees were predominantly 117 Jack pine (Pinus banksiana) and Red pine (Pinus resinosa), but also included white pine (Pinus strobus), 118 white spruce (Picea glauca) and green alder (Alnus viridis) (VETAC, 2020). 120 Figure 2. Map of the City of Greater Sudbury, Ontario, Canada including sixteen study sites, the outlines 121 for the barren and semi-barren regions, and the location of the three main smelters. 123 2.2 Site Selection 124 Sixteen sites varying in age were established within the barrens of the Coniston and Copper Cliff 125 smelters (Figure 2). The barrens were defined as areas that were devoid of plant life and had soil pH 126 below 4.3 before regreening efforts took place (Lautenbach et al. 1995). All sites were limed, fertilized, Electronic copy available at: https://ssrn.com/abstract=3954017 127 seeded, and planted with either red pine and/or jack pine. Planting dates ranged from 1979 to 2006, and 128 sites that only received planting at one time were specifically selected. Sites were split into two erosion 129 classes; highly eroded sites had greater than 30% bedrock exposure and typically had very little soil. 130 Sites with less than 10% bedrock exposure and deeper soils were considered “stable”. Six eroded sites 131 and ten stable sites were included in the study sites. Each site was 30 m x 30 m in size within a larger 132 forested area with similar characteristics. 134 2.3 Field sampling 135 Field sampling was conducted in June 2021. At each site, diameter at breast height (dbh) and 136 tree species were recorded for each tree. Trees were categorized as woody plants with a dbh larger than 137 2.5 cm. Soil depth was measured at nine points at each site using a soil auger. Forest floor depths were 138 taken at nine points by separating the forest floor from mineral soil using a knife. Elevation was 139 measured at each site using a Garmin eTrex Legend H GPS. Slope was measured at each site using a 140 Suunto PM-5/1520 clinometer. Tree cores were taken from three trees in each plot to verify age and 141 used for wood nutrient analysis. Two cores were taken from each sampled tree, one at breast height 142 and the other 20 cm below at a 90⁰ angle from the entrance point of the initial core. Cores were stored 143 in plastic straws with the ends taped. Bark samples were taken from these same trees at breast height 144 by scraping bark off the tree using a knife and composited from the three trees in a polyethylene bag. 145 Foliage samples were taken by hand from highest possible branch for 3 – 5 individual trees and 146 composited into paper bags. Current year needles were targeted for foliage sampling. Most sites were 147 monocultures of jack pine or red pine; cores, bark and foliage samples were only completed for the 148 dominant species at each site. On the occasion that the stands had a more heterogenous species 149 composition, samples were taken for any species that comprised greater than 33% of individuals per Electronic copy available at: https://ssrn.com/abstract=3954017 150 site. All vegetation (cores, bark, and foliage) samples were stored in a refrigerator (4 ⁰C) upon return 151 from the field. Vegetation samples were only collected for eleven of the sixteen study sites. 153 2.4 Chemical analysis 154 All vegetation samples were milled to a fine powder using a coffee grinder. Milled samples were 155 analyzed for Ca, Mg, K and P by digesting 0.2 g of sample in 2.5 mL of concentrated nitric acid (trace 156 grade 70% nitric acid (HNO )) at 100 ⁰C for eight hours followed by a digestion at room temperature for 157 eight hours. The digested material was filtered through a 0.45 µm pore filter paper and diluted using B- 158 pure until total volume reached 25 mL. The digest samples were then diluted 1:10 using B-pure and 159 analyzed using a Perkin Elmer Inductively Ion Coupled Optical Emission Spectrometer (7000 DV). Carbon 160 and N values were determined for the milled vegetation samples using an Elementar Macro CNS 161 analyzer. Quality control was confirmed by running all samples in triplicate and including experimental 162 blanks as well as a NIST-1515 apple leaf standard every 25 samples. 164 2.5 Aboveground biomass and nutrient pools 165 Aboveground biomass was calculated for each site using dbh-based allometric equations from 166 Lambert et al. (2005). Aboveground nutrient (Ca, Mg, N, P, K, and C) pools were determined using 167 measured nutrient concentrations for each biomass component (bark, foliage, wood + branches) and 168 the mass of each biomass component from the allometric equations. Tree core chemistry was used for 169 both wood and branches biomass components. If a site contained multiple species with greater than 170 33% individuals than chemistry for both red and jack pine were averaged. Aboveground biomass pools 171 were calculated for all sixteen study sites, aboveground nutrient pools were calculated for the eleven 172 sites where chemical analyses on vegetation samples were conducted. Aboveground biomass and Electronic copy available at: https://ssrn.com/abstract=3954017 173 nutrient accumulation rates were determined by dividing AGB or nutrient pool size by years since tree 174 planting. 176 2.6 Landsat image processing 177 The Landsat image was acquired from USGS Landsat 8 OLI program from July 9, 2020 (path: 19 178 and row: 28) level 1T (L1T), an image was selected that had less than 5% cloud cover. Landsat Level-1 179 products are geometrically corrected by orthorectification of ground control points and digital elevation 180 models (USGS 2019). The image was radiometrically and atmospherically corrected using the semi- 181 classification plugin based on the dark object subtraction (DOS1) technique (Chavez 1996) on Quantam 182 GIS (QGIS) Version 3.2.1 (QGIS.org 2020). 184 2.7 Random forest model development and application 185 A RF model was developed using Landsat 8 OLI bands 2 – 7, and ten vegetation indices (VIs) 186 (Table 1) at 35 sites within the City of Greater Sudbury barren areas where AGB was measured. Each 187 remote sensing variable included in model development has previously been shown to reflect various 188 components of AGB (Chenge & Osho, 2018; López-Serrano et al. 2020; Zhang et al. 2019). The 35 sites 189 used for model development included: the sixteen study sites, ten sites planted with red and/or jack 190 pine from Rumney et al. (2021), and nine sites that had an AGB of zero (three areas with bare exposed 191 bedrock, three roads, and three lakes). The study sites and zero AGB sites were 30 m by 30 m squares 192 while the Rumney et al. (2021) sites were circular plots with a diameter of 35.68 m. 194 To assess AGB and AGB nutrient pools for the regreening areas the RF model was applied to the 195 19,649 ha where red and/or jack pine had been planted across the barren areas of the City of Greater 196 Sudbury. Nutrient pools were determined for 7,216 ha of the barren areas where red and/or jack pine Electronic copy available at: https://ssrn.com/abstract=3954017 197 were planted and liming, and fertilizing took place. Magnesium, K, and C AGB concentrations were 198 determined by averaging the AGB nutrient concentrations from the study sites. Calcium, P, and N AGB 199 concentrations were determined using a linear curve based on the results from the study sites where 200 nutrient concentration varied with time since tree planting. Location and dosage rate of lime and 201 fertilizer applications were available from the City of Greater Sudbury regreening data (City of Greater 202 Sudbury 2020). 204 Table 1. Vegetation indices included in random forest to develop predictive model for aboveground 205 biomass using remotely sensed data. Landsat 8 OLI bands depicted using b. L is soil brightness correction 206 factor and its value is 0.5. Vegetation Index Equation Reference Brightness index (BI) Baig et al. (2014) 𝑏2 ∗ 0.3029 + 𝑏3 ∗ 0.2786 + 𝑏4 ∗ 0.4733 + 𝑏5 ∗ 0.5599 + 𝑏6 ∗ 0.508 + 𝑏7 ∗ 0.1872 Green normalized difference vegetation index (GNDVI) (𝑏5 ‒ 𝑏3)/(𝑏5 + 𝑏3) Gitelson et al. (1996) Greenness vegetation index (GVI) 𝑏2 ∗‒ 0.2941 + 𝑏3 ∗‒ 0.243 + 𝑏4 ∗‒ 0.5424 + 𝑏5 ∗ 0.7276 + 𝑏6 ∗ 0.0713 + 𝑏7 ∗‒ 0.1608 Baig et al. (2014) Land surface water index (LSWI) Xiao et al. (2004) (𝑏5 ‒ 𝑏6)/(𝑏5 + 𝑏6) Modified soil adjusted vegetation index (MSAVI) Qi et al. (1994) ((𝑏5 ‒ 𝑏4)/(𝑏5 + 𝑏4)) ∗ (1 + 𝐿) Normalized difference infrared index (NDII) (𝑏5 ‒ 𝑏6)/(𝑏5 + 𝑏6) Kimes et al. (1981) Normalized difference infrared index with SWIR2 (NDII7) (𝑏5 ‒ 𝑏7)/(𝑏5 + 𝑏7) Kimes et al. (1981) Normalized difference vegetation index (NDVI) (𝑏5 ‒ 𝑏4)/(𝑏5 + 𝑏4) USGS, (2016) Normalized difference water index (NDWI) Gao, (1996) (𝑏3 ‒ 𝑏6)/(𝑏3 + 𝑏6) Wetness index (WI) 𝑏2 ∗ 0.1511 + 𝑏3 ∗ 0.1973 + 𝑏4 ∗ 0.3283 + 𝑏5 ∗ 0.3407 + 𝑏6 ∗‒ 0.7117 + 𝑏7 ∗‒ 0.4559 Baig et al. (2014) 208 2.8 Statistical analysis 209 The relationship between AGB (at the hectare and stem scale), and AGB nutrient concentrations 210 with time since tree planting was assessed using regression coefficient (R ) values from linear regression. 211 Normality was tested for using the Lilliefors corrected Kolmogorov-Smirnov test, variables were 212 considered normally distributed if p was greater than 0.05. Differences between erosion classes in AGB 213 and nutrient accumulation rates were compared using an independent two-group Mann-Whitney- 214 Wilcoxon test; significant differences were determined if p was less than 0.05. The AGB predictive model 215 was developed using RF regression for seventeen remotely sensed predictor variables and AGB 216 measurements from 35 sites. Random forest was conducted through the Forest-based Classification and Electronic copy available at: https://ssrn.com/abstract=3954017 217 Regression function in ArcGIS Pro version 2.8.2. (ArcGIS Pro, 2021). The training of the RF model was 218 conducted on 90% of the available data with 10% being held back for validation. The selected RF model 219 included 1500 decision trees, 5 randomly selected variables, and 4 nodes per tree. The RF model was 220 evaluated on the training data set using residual mean squared error (RMSE) and regression coefficient 2 2 221 (R ) values and validated using the R value from the validation data set as well as out-of-bag RMSE. All 222 statistical analysis was done on R version 4.0.2 (RStudio Team 2020). All maps were created using 223 Quantam GIS (QGIS) Version 3.2.1 (QGIS.org 2020). 225 3. Results 226 Total stand age ranged from 14 to 41 years old with a median age of 27 years old (Table 2). 227 Median eroded site age was 27 years old while median stable site age was 25 years old. Fifty percent of 228 the total sites were dominated by jack pine, 25% by red pine and 25% were a mix of red and jack pine 229 (Table 2). Stem density ranged from 578 to 2,378 stems/ha with a median stem density of 1,660 230 stems/ha. Eroded sites had a median stem density of 1,144 stems/ha while stable sites had a stem 231 density of 1,668 stems/ha. Average tree dbh ranged from 5.1 cm to 16.0 cm with a median of 9.5 cm. 232 Both eroded and stable sites had a median tree dbh of 9.5 cm. Soil depth was a median of 25.06 cm 233 deep for stable sites and 6.00 cm for eroded sites. 235 Table 2. Site characteristics of sixteen study sites within the barren areas of the City of Greater Sudbury, 236 Ontario, Canada used to develop relationship between aboveground biomass and stand age. Tree Forest Mineral density floor soil Age Erosion Elevation Dominant (stems/ DBH depth depth Site Easting Northing (years) class (m) Slope (⁰) species ha) (cm) (cm) (cm) CC1 498390 5148618 24 Stable 309 4 Jack pine 1,822 14.3 4.58 15.06 CC2 498513 5148661 32 Eroded 300 15 Jack pine 1,866 8.1 1.07 3.56 CC3 498375 5150759 23 Stable 298 0 Red pine 1,644 10.9 1.32 25.87 CC10 495810 5143521 37 Stable 215 8 Red pine 1,655 14.7 5.00 23.33 Electronic copy available at: https://ssrn.com/abstract=3954017 CON1 515267 5148285 34 Stable 270 15 Jack pine 1,078 14.3 4.35 32.62 CON3 513804 5148482 40 Stable 211 8 Jack pine 1,744 8.1 1.78 14.31 CON4 513724 5148484 32 Eroded 258 13 Jack pine 1,666 8.8 0.88 4.94 CON5 512289 5149244 28 Stable 268 3 Mixed 1,666 7.4 4.25 22.94 CON6 512320 5149231 28 Eroded 281 6 Red pine 889 13.2 1.63 10.06 CON7 511116 5147545 21 Stable 241 0 Mixed 2,378 7.0 1.37 46.40 CON9 509020 5149553 15 Eroded 304 6 Jack pine 1,211 5.1 1.03 6.00 CON10 509140 5149646 15 Stable 290 14 Mixed 1,500 6.5 0.92 25.06 CON12 513979 5148306 42 Eroded 278 17 Jack pine 578 16.0 3.50 6.22 CON15 508894 5149251 24 Stable 215 14 Mixed 1,711 7.7 5.50 23.33 CON16 511895 5149249 34 Stable 278 2 Red pine 2,178 14.5 7.00 24.33 D1 509127 5145262 22 Eroded 295 2 Jack pine 1,078 10.2 1.33 4.87 Electronic copy available at: https://ssrn.com/abstract=3954017 238 239 Figure 3. Relationship between aboveground biomass (AGB) and time since planted for both highly 240 eroded sites with >30% rock cover and stable sites (soil cover >80%). Figure 3A depicts relationship at 241 the hectare scale; figure 3B is per individual stem. Dashed lines represent 95% confidence intervals. 242 Regression coefficient (R ) for relationship between AGB and age for both erosion classes is included. 244 Aboveground biomass (Mg/ha) had a significant positive linear relationship with time since tree 2 2 245 planting at the site scale for both the eroded (R =0.97, p<0.01) and stable sites (R =0.91, p<0.01) (Figure Electronic copy available at: https://ssrn.com/abstract=3954017 246 3A). Stable sites (2.78 ± 1.05 Mg/ha/year) had a significantly larger (p<0.01) AGB growth rate compared 247 with eroded sites (1.55 ± 0.38 Mg/ha/year). Aboveground biomass (kg/stem) also had a significant 248 positive linear relationship with time since tree planting at the stem scale for eroded (R =0.78) and 249 stable sites (R =0.82), however, AGB growth rates per stem were not significantly different (p=0.30) 250 between eroded (1.34 ± 0.71 kg/stem/year) and stable sites (1.59 ± 0.85 kg/stem/year) (Figure 3B). 252 Table 3. Aboveground biomass nutrient concentrations and accumulation rates (at both the hectare and 253 stem scale). Asterix denotes significant difference in nutrient concentration, or accumulation rate 254 between eroded and stable erosion classes (p<0.05). Nutrient concentrations Accumulation Rate Accumulation Rate Nutrient Erosion Class (g/kg) (kg/ha/year) (g/stem/year) Calcium Eroded 1.44 ± 0.51 1.80 ± 0.29 1.49 ± 0.34 Calcium Stable 1.10 ± 0.21 3.06 ± 0.76 1.61 ± 0.88 Magnesium Eroded 0.31 ± 0.08 0.48 ± 0.16 0.26 ± 0.20 Magnesium Stable 0.26 ± 0.05 0.74 ± 0.23 0.41 ± 0.19 Nitrogen Eroded 2.07 ± 0.35 2.77 ± 0.69* 2.34 ± 0.76 Nitrogen Stable 2.22 ± 0.43 5.18 ± 1.77* 2.90 ± 2.10 Phosphorous Eroded 0.14 ± 0.05 0.18 ± 0.04* 0.13 ± 0.07 Phosphorous Stable 0.15 ± 0.06 0.37 ± 0.12* 0.24 ± 0.07 Potassium Eroded 0.61 ± 0.17 0.75 ± 0.16* 0.67 ± 0.18 Potassium Stable 0.67 ± 0.15 1.51 ± 0.38* 0.90 ± 0.34 Carbon Eroded 483.53 ± 3.16 726.67 ± 199.51* 574.17 ± 188.70 Carbon Stable 485.62 ± 6.36 1279.70 ± 402.59* 669.08 ± 475.10 258 Aboveground nutrient concentrations were not significantly different between eroded or stable 259 sites (p>0.05) (Table 3). Accumulation rates of N (p<0.01), P (p<0.01), K (p<0.01), and C (p<0.05) within 260 AGB were significantly higher for stable versus eroded sites at the hectare scale. However, at the stem 261 scale there was no significant difference between nutrient accumulation rates and site erosion class 262 (p>0.05) (Table 3). There was a significant negative linear relationship between AGB concentration of Ca 2 2 2 263 (R =-0.41, p<0.05) (Figure 4A), N (R =0.-60, p<0.01) (Figure 4B), and P (R =-0.62, p<0.01) (Figure 4C) and Electronic copy available at: https://ssrn.com/abstract=3954017 264 time since tree planting and no relationship between concentration of Mg, K, or C in AGB and time since 265 tree planting (p>0.05) (Table 4). There was a significant positive linear relationship between time since 266 tree planting and AGB pools of all measured nutrients (p<0.05) (Table 4). Electronic copy available at: https://ssrn.com/abstract=3954017 267 268 Figure 4. Relationship between aboveground nutrient concentrations and time since tree planting for 269 calcium (Ca) (Figure 4A), nitrogen (N) (Figure 4B), and phosphorous (P) (Figure 4C). Dashed lines 270 represent 95% confidence intervals. Regression coefficient (R ) is included. Electronic copy available at: https://ssrn.com/abstract=3954017 2 272 Table 4. Regression coefficients (R ) for the relationships between aboveground biomass (AGB) nutrient 273 concentrations and time since tree planting, and the relationship between AGB nutrient pool sizes and 274 time since tree planting. Asterix denotes significant linear relationship (p<0.05). AGB nutrient AGB nutrient pool concentration (g/kg) size (g/stem) and and time since tree time since tree 2 2 Nutrient planting (R ) planting (R ) Calcium -0.41* 0.78* Magnesium -0.10 0.80* Nitrogen -0.62* 0.67* Phosphorous -0.60* 0.78* Potassium 0.24 0.82* Carbon -0.01 0.72* 278 Figure 5. Predicted aboveground biomass (AGB) values compared with measured AGB for 35 sites where 279 AGB was calculated. Black solid line represents 1:1 line. Solid green line represents regression line. 280 Dashed green lines represent 95% confidence intervals. Relationship between measured and predicted 281 AGB values assessed using root mean square error (RMSE) along with regression coefficient (R ). Electronic copy available at: https://ssrn.com/abstract=3954017 282 283 Aboveground biomass predictions from the RF model had a R of 0.91 and RMSE of 13.6 Mg/ha 284 when compared to AGB measurements from the training data set. Validation of the RF model revealed a 285 R of 0.91 when compared to AGB measurements of the validation data set and an out-of-bag RMSE of 286 31.35 Mg/ha. Total AGB predicted using the RF model for red and jack pine planted sites (19,649 ha) of 287 the City of Greater Sudbury is 1, 144, 588 Mg (550, 547 Mg C) (Figure 6). The mass of applied Ca and Mg 288 (in the form of crushed limestone) were 20 – 100 times greater than the mass of applied N, P, or K (in 289 the form of 6-24-24 fertilizer) for the 5,848 ha of limed, fertilized, and reforested land (Table 5). The 290 mass of N stored in AGB was over six times as high as the mass applied. The proportion of P and K stored 291 within AGB compared to applied nutrients was also substantial (Table 5). Electronic copy available at: https://ssrn.com/abstract=3954017 294 Figure 6. Aboveground biomass for red pine and jack pine planting polygons across the City of Greater 295 Sudbury, Ontario, Canada. 297 Table 5. Comparison of applied nutrients (applied as either lime or fertilizer) with nutrients stored in 298 aboveground biomass (AGB). Aboveground biomass nutrient pools are for all sites in the City of Greater 299 Sudbury that were limed, fertilized, and planted with Red and/or Jack pine (5,848 ha). Proportion of applied Nutrient concentration nutrients stored in Nutrients Applied (Mg) (g/kg) AGB (%) Calcium 12,707 1.25 ± 0.41 3.74 Magnesium 7707 0.28 ± 0.07 1.17 Nitrogen 126 2.16 ± 0.40 614.41 Phosphorous 239 0.14 ± 0.06 24.84 Potassium 454 0.64 ± 0.16 45.15 302 4. Discussion 303 Greater AGB pools at stable versus eroded sites was initially assumed to be due to deeper soils 304 leading to increased stem growth, as both jack pine and red pine have been found to be more 305 productive in deeper soils (Rudolph & Laidly 1990; Rudolph 1990). However, this was not the case as 306 individual stem growth did not differ between the two erosion classes and differences in hectare scale 307 AGB totals is driven primarily by stem density. Initial planting densities reported by the City of Greater 308 Sudbury (2020) were similar to measured stem density at the study sites, suggesting that the higher 309 stem density on stable sites is likely due to increased planting density as opposed to other factors such 310 as increased survivorship. Previous studies assessing C accumulation rates of jack and red pine 311 plantations in northern Ontario, Canada (Hunt et al. 2010), and Manitoba, Canada (Park, 2015), 312 converted AGB growth to C accumulation rates assuming C comprised 48% of AGB for Red pine and 50% 313 for Jack pine (Hunt et al. 2010; Park 2015). The individual stem growth rates from both Hunt et al. (2010) 314 (1.33 ± 0.33 kg/stem/year) and Park (2015) (0.92 ± 0.53 kg/stem/year) were very similar to values Electronic copy available at: https://ssrn.com/abstract=3954017 315 measured in the current study. Despite the history of highly acidified soils, high total metal 316 concentrations the trees in the regreening sites of the City of Greater Sudbury are growing at a similar 317 rate to those of jack and red pine plantations in relatively undisturbed areas regardless of erosion class. 319 Aboveground Ca, Mg, N, P, and K concentrations did not differ between eroded and stable sites 320 and median concentrations were within a standard deviation of healthy red pine nutrient 321 concentrations included in the United States Department of Agriculture Tree Chemistry Database (TCD) 322 (Pardo et al. 2005) (Table 6). Carbon concentration for wood at the study sites was 48.1 ± 0.6%, firmly in 323 between the Martin et al. (2018) values for boreal (46.8 ± 0.6%) and temperate conifers (50.1 ± 0.4%). 324 While the median AGB nutrient concentrations for the study sites are similar to the healthy values from 325 the TCD, concentrations of Ca, P, and N are decreasing with increasing time since tree planting. A 326 decrease in total AGB concentration of Ca, N, and P with increasing stand age has been documented 327 previously as these nutrients are stored in high concentrations in foliage which accumulates mass very 328 quickly in young trees as opposed to the slower growing, more nutrient poor wood (MacLean and Wein, 329 1977). The magnitude of the decrease for Ca and P at the study sites is cause for some concern as trees 330 at the two oldest study sites (CON1 and CON3) have bark and wood Ca and P concentrations well below 331 TCD values. Conversely, despite the decreasing concentration of N with time since tree planting, AGB N 332 concentration at CON1 and CON3 still exceed TCD values. 334 Table 6. Nutrient concentrations for each biomass component. Both the measured concentrations from 335 the study sites and the concentrations listed for Red pine from the Tree Chemistry Database (TCD) 336 (Pardo et al. 2005) are included. Bark Foliage Wood Study sites Study sites Study sites Nutrient (g/kg) TCD (g/kg) (g/kg) TCD (g/kg) (g/kg) TCD (g/kg) Electronic copy available at: https://ssrn.com/abstract=3954017 Ca 3.78 ± 2.17 7.65 ± 3.91 3.67 ± 1.81 4.17 ± 1.60 0.80 ± 0.15 1.09 ± 0.35 Mg 0.31 ± 0.14 0.46 ± 0.13 1.27 ± 0.62 0.86 ± 0.34 0.19 ± 0.05 0.19 ± 0.12 N 4.82 ± 1.28 3.10 ± 0.38 10.46 ± 1.16 11.49 ± 1.30 1.24 ± 0.40 0.82 ± 0.11 P 0.33 ± 0.20 0.44 ± 0.08 1.27 ± 0.33 1.28 ± 0.16 0.03 ± 0.03 0.08 ± 0.02 K 0.47 ± 0.47 0.88 ± 0.27 4.36 ± 1.59 3.67 ± 1.00 0.35 ± 0.12 0.24 ± 0.09 C 525.92 ± 26.15 Not available 492.50 ± 15.73 Not available 480.59 ± 5.90 Not available 338 Aboveground nutrient accumulation rates do not differ between erosion classes once stem 339 density is controlled for. Accumulation rates of Ca, Mg, N, P, K, and C for the study sites, increase over 340 time with a pattern similar to AGB growth and closely reflect values from coniferous plantations in 341 similar climates (Alban et al. 1978; Hunt et al., 2010; MacLean and Wein, 1977; Morrison, 1973; Park, 342 2015). Morrison (1973) assessed nutrient accumulation in a 30-year-old jack pine plantation in northern 343 Ontario, Canada at both high- and low-quality sites defined by biomass yield. The accumulation rates of 344 AGB Ca, Mg, P, and K at the City of Greater Sudbury regreening study sites closely reflected stands from 345 Morrison (1973) grown on poorer quality soils, while AGB N accumulation rates are more similar to the 346 higher quality sites. The similarity of the AGB growth rates, nutrient concentrations, and nutrient 347 accumulation rates to plantations unimpacted by mining and smelting practices suggest soil 348 amelioration has generally been a success in the City of Greater Sudbury barren areas. Soils in this area 349 were once inhospitable and now appear to support tree growth and nutrient accumulation similar to 350 poor quality but natural soils that are native to the southern portion of the Precambrian Shield. 352 A RF model incorporating ten VIs and Landsat 8 OLI bands 2 – 7 proved to be a successful 353 predictor of AGB for the study sites with higher R values and lower RMSE compared to studies using 354 similar methods (Gao et al. 2018; Jiang et al. 2021; López-Serrano et al. 2020; Zhang et al. 2019). The 355 stronger relationship and lower error for the model in the current study could be attributed to the 356 regreening forests being a low complexity stands, often a monoculture with limited understory; it has 357 been reported that models using VIs have decreased accuracy with high complexity forests (Lu et al. Electronic copy available at: https://ssrn.com/abstract=3954017 358 2016). An additional reason for the low RMSE of AGB estimates at the study sites, is the low AGB of 359 these forests, Landsat regression-based models have been found to become saturated at AGB values of 360 150 Mg/ha (Lu 2005). 362 The total AGB for red and jack pine stands across 19,649 ha within the City of Greater Sudbury is 363 1, 144, 588 Mg. Using an average of the eroded and stable site growth-curves the average AGB of 58.25 364 Mg/ha implies that the average site is 27.3 years old. As the remediated stands continue to age, we 365 expect AGB C to increase as jack pine accumulate biomass up to 100 years old and Red pines can 366 accumulate biomass until up to 300 years old (Rudolph & Laidly 1990; Rudolph 1990). The total AGB C 367 pool of 550, 547 Mg C is the equivalent to C emissions from 319, 158 cars over one year (3.67% of the 368 registered vehicles in the Province of Ontario) (NRCAN 2014). Applying the federal governments price 369 for C at $50 CAD per Mg, the AGB C pools of the regreening sites of the City of Greater Sudbury is equal 370 to $27.53 million CAD since the restoration project began in 1978. This total does not include any other 371 C pools such as mineral soil, forest floor, shrub, and herb pools. Preston et al. (2020) found that while 372 forest floor C increased with stand age, the mineral soil C pool’s relationship with stand age was unclear 373 and further work on the City of Greater Sudbury regreening sites suggests that soil C is decreasing with 374 stand age (Levasseur et al. unpublished). The current study’s AGB C projections only include sites 375 planted with jack and red pine, while Jack and Red pine are the most commonly planted tree species, 376 there are regions planted with other conifers and deciduous species that are not included in this total 377 (VETAC, 2020). 379 A concern for long-term forest growth on degraded landscapes is lack of nutrients. To overcome 380 the low nutrient concentrations of the City of Greater Sudbury’s barren soils, the regreening program 381 applied massive quantities of nutrients (particularly Ca and Mg) to the landscape (Winterhalder 1996). Electronic copy available at: https://ssrn.com/abstract=3954017 382 Only 4% of applied Ca and 1% of applied Mg were stored within AGB, this is most likely due to large- 383 scale erosion removing Ca and Mg from the landscape. A recent study conducted on the same sites as 384 the current study found that large quantities of applied Ca and Mg were unaccounted for in nutrient 385 budgets (Kellaway et al. 2021). High percentages of P, and K were accumulated within AGB which could 386 also be an indication of nutrient limitation. Phosphorous has often been reported to be the primary 387 limiting nutrient within the barrens of the City of Greater Sudbury (Beckett & Negusanti 1990; 388 Lautenbach et al. 1995; Munford et al. 2021). The limited soil organic matter, slow decomposition rates 389 caused a limited source of available P and the low soil pH caused large proportions of available P to be 390 bound to inorganic complexes (Marschner 1991). Previous studies on the City of Greater Sudbury 391 barrens found that when liming and fertilizer has been applied, vegetation rapidly accumulates P in 392 biomass (Winterhalder 1996). Little work has been done on K limitation in Sudbury; however, liming and 393 fertilizing has been found to either not impact or decrease bioavailable K in soils within the City of 394 Greater Sudbury (Nkongolo et al. 2013; SARA Group 2009). Despite the limited bioavailability of K in 395 soils, Kellaway et al. (2021) found that the equivalent of 95% of applied K at the study sites was stored in 396 biomass. 398 Nitrogen pools stored within AGB were found to be over six times as high as the amount applied 399 in fertilizer, implying trees were accumulating N from additional sources outside of fertilizer. The SARA 400 Group (2009) found that total N concentrations in forest soils of the City of Greater Sudbury barrens 401 were lower than those of reference forests, however, residual organic matter in the forest soils of the 402 barrens has been proposed to be a sufficiently large pool to sustain vegetation communities 403 (Winterhalder 1995). The pre-restoration forest soil N pools may make up the difference between N 404 applied as fertilizer and N stored in AGB. Additionally, N-fixing species such as Lotus corniculatus (Bird’s- 405 foot trefoil), Trifolium hybridum (Alsike clover) were planted widely during restoration, and the native Electronic copy available at: https://ssrn.com/abstract=3954017 406 Comptonia peregrina (Sweet fern), and Pleurozium sp. (Feather moss) have become common in the 407 recovering jack pine stands (Lautenbach et al. 1995). Limited recent work has been done on changes in 408 forest soil N pools within the City of Greater Sudbury, however, a study by Meyer-Jacob et al. (2020) 409 found that organic N has been increasing in the region’s lakes from 1984 – 2020. 411 While trees within the remediated sites appear to be accumulating N from additional sources 412 and have AGB N concentrations higher than healthy conifers this is not apparent for Ca, P, and K. 413 Calcium and P are of particular concern as both nutrients have declining concentrations with time since 414 tree planting and have concentrations lower than healthy conifers in the oldest study sites. As the 415 average age for the regreening sites is 27.3 years, existing Ca, P, and K pools, provided from fertilizer and 416 limestone may be depleted as these stands continue to age. An additional concern for future forest 417 growth is total metal concentrations in the regreening area, such as Ni and Cu, remain high in mineral 418 soils; however, bioavailable concentrations of these metals are very low (reported between 0.1 – 1% of 419 total metal concentrations) (Kellaway et al. 2021; Nkongolo et al. 2013). While the high total metal 420 concentrations do not appear to be affecting tree growth up to this point in the regreening program, 421 Kellaway et al. (2021) found that metal partitioning at the study sites was driven more by pH versus OM, 422 indicating that as effects of liming on pH diminish, metals may become more bioavailable potentially 423 impacting tree growth. 425 5. Conclusions 426 The City of Greater Sudbury regreening efforts are regarded as one of the world’s largest 427 restoration projects and have led to over one million tonnes of AGB (over 500, 000 tonnes of C) 428 establishing in the 19, 649 ha of regreening area over 40 years. Within the regreening area, individual 429 trees are growing and accumulating aboveground nutrients at rates similar to jack and red pine Electronic copy available at: https://ssrn.com/abstract=3954017 430 plantations grown in areas unimpacted by mining and smelting. Aboveground nutrient concentrations 431 exceeding TCD values (particularly N) similarly suggest that up to 40-years post restoration the 432 amendments applied to the City of Greater Sudbury regreening areas appear to be sufficient to support 433 healthy tree growth. However, some cause for concern going forward is aboveground P concentrations 434 are decreasing as trees age and 40-year old study sites already have aboveground P concentrations 435 below “healthy” tree values. Phosphorous limitation has been previously documented in the regreening 436 areas and future studies should continue to investigate potential future nutrient limitation. 438 Acknowledgements 439 This study is part of the Landscape Carbon Accumulation through Reductions in Emissions (L- 440 CARE) project and is funded by Ontario Centers for Excellence, Natural Sciences and Engineering 441 Research Council of Canada, Vale Canada Limited, Sudbury Integrated Nickel Operations and the City of 442 Greater Sudbury. The authors would like to thank the countless volunteers responsible for tree planting, 443 liming, and fertilizing the City of Greater Sudbury barrens. The authors would also like to thank the rest 444 of the L-CARE team as well as Kevin Adkinson, Kaylen Foley, Spencer Gilbert-Parkes, William Humphrey, 445 Edward Kellaway, Kimber Munford, and Jodi Newman for technical support throughout this study. The 446 authors declare that they have no conflict of interests. Electronic copy available at: https://ssrn.com/abstract=3954017 447 Literature cited 448 Alban DH, Perala DA, and Schlaegel BE (1978) Biomass and nutrient distribution in aspen, pine and 449 spruce stands on the same soil type in Minnesota. Canadian Journal of Forest Research 8: 290 – 450 299 451 ArcGIS Pro (2021) ESRI https://www.esri.com/en-us/arcgis/products/arcgis-pro/overview 452 Bastin JF, Finegold Y, Garcia C, Mollicone D, Rezende M, Routh D, Crowther, T. W. et al (2019). The 453 global tree restoration potential. Science 365: 76 – 79 454 Baig MHA, Zhang L, Shuai T, Tong Q (2014) Derivation of a tasselled cap transformation based on 455 Landsat 8 at-satellite reflectance. Remote Sensing Letters 5: 423 – 431 456 Beckett PJ, Negusanti J (1990) Using land reclamation practices to improve tree condition in the Sudbury 457 smelting area, Ontario, Canada. Proceedings of the 1990 Mining and Reclamation Conference 458 and Exhibition West Virginia University, Morgantown, West Virginia, USA 459 Chavez PS (1996) Image-based atmospheric corrections – revisited and improved. Photogrammetric 460 Engineering & Remote Sensing 62: 1025 – 1036 461 Chenge IB, Osho JSA (2018) Mapping tree aboveground biomass and carbon in Omo Forest Reserve 462 Nigeria using Landsat 8 OLI data. Southern Forests 80: 341 – 350 463 City of Greater Sudbury (2020) Regreening program https://www.greatersudbury.ca/live/environment- 464 and-sustainability1/regreening-program/ 465 Dieleman CM, Rogers BM, Potter S, Veraverbeke S, Johnstone JF, Laflamme J, Solvik K, Walker XJ, Mack 466 MC, Turetsky MR (2020) Wildfire combustion and carbon stocks in the southern Canadian boreal 467 forest: Implications for a warming world. Global Change Biology 26: 6062 – 6079 468 Environment Canada (2019) Canadian climate normals 1981 – 2010 station data 469 https://climate.weather.gc.ca/climate_normals/results_1981_2010_e.html Electronic copy available at: https://ssrn.com/abstract=3954017 470 Friedlingstein P, Allen M, Canadell JG, Peters GP, Seneviratne SI (2019) Comment on “The global tree 471 restoration potential”. Science 366: 1 – 2 472 Gao BC (1996) NDWI—A normalized difference water index for remote sensing of vegetation liquid 473 water from space. Remote sensing of environment 58: 257 – 266 474 Gao Y, Lu D, Li G, Wang G, Chen Q, Liu L, Li D (2018) Comparative analysis of modeling algorithms for 475 forest aboveground biomass estimates in a subtropical region. Remote Sensing 10: 627 476 Gitelson AA, Kaufman YJ, Merzlyak MN (1996) Use of a green channel in remote sensing of global 477 vegetation from EOS-MODIS. Remote sensing of Environment 58: 289 – 298 478 Gordon WS, Jackson RB (2000) Nutrient concentrations in fine roots. Ecology 81: 275 – 280 479 Gorgolewski A, Rudz P, Jones T, Basiliko N, Caspersen J (2020) Assessing coarse woody debris nutrient 480 dynamics in managed northern hardwood forests using a matrix transition model. Ecosystems 481 23: 541 – 554 482 Gunn J, Keller, W Negusanti J, Potvin R, Beckett P, Winterhalder K (1995) Ecosystem recovery after 483 emission reductions: Sudbury, Canada. Water, Air, & Soil Pollution 85: 1783 – 1788 484 Holl KD, Brancalion PHS (2020) Tree planting is not a simple solution. Science 368: 580 – 581 485 Hooker TB, Compton JE (2003) Forest ecosystem carbon and nitrogen accumulation during the first 486 century after agricultural abandonment. Ecological Applications 13: 299 – 313 487 Hunt SL, Gordon AM, Morris DM (2010) Carbon stocks in managed conifer forests in northern Ontario, 488 Canada. Silva Fennica 44: 563 – 582 489 International Union for the Conservation of Nature (IUCN) (2020) IUCN Annual 2019 Report. IUCN, 490 Gland, Switzerland. 491 Jiang F, Kutia M, Ma K, Chen S, Long J, and Sun H (2021). Estimating the aboveground biomass of 492 coniferous forest in Northeast China using spectral variables, land surface temperature and soil 493 moisture. Science of The Total Environment 785: 147335. Electronic copy available at: https://ssrn.com/abstract=3954017 494 Johnson DW, Todd DE (1998) Harvesting effects on long-term changes in nutrient pools of mixed oak 495 forest. Soil Science Society of America Journal 62: 1725 – 1735 496 Kellaway EJ, Eimers MC, and Watmough SA (2021). Liming legacy effects associated with the world's 497 largest soil liming and regreening program in Sudbury, Ontario, Canada. Science of The Total 498 Environment: 150321. 499 Kimes DS, Markham BL, Tucker CJ, & McMurtrey III JE (1981) Temporal relationships between spectral 500 response and agronomic variables of a corn canopy. Remote Sensing of Environment 11: 401 – 501 411 502 Lambert M-C, Ung C-H, Raulier F, (2005) Canadian national tree aboveground biomass equations 503 Canadian Journal of Forest Research 35: 1996 – 2018 504 Lautenbach WE, Miller J, Beckett PJ, Negusanti JJ, Winterhalder K (1995) Municipal land restoration 505 program: the regreening process. Pages 109 – 122 In: Gunn J (ed) Restoration and recovery of an 506 industrial region. Springer, New York, USA 507 Likens GE, Driscoll CT, Busco DC, Siccama TG, Johnson CE, Lovett GM, Fahey TJ, Reiners WA, Ryan DF, 508 Martin CW, Bailey SW (1998) The biogeochemistry of calcium at Hubbard Brook. 509 Biogeochemistry 41: 89 – 173 510 Likens GE, Driscoll CT, Busco DC, Siccama TG, Johnson CE, Lovett GM, Ryan DF, Fahey TJ, Reiners WA 511 (1994) The biogeochemistry of potassium at Hubbard Brook. Biogeochemistry 25: 61 – 125 512 López-Serrano PM, Cárdenas Domínguez JL, Corral-Rivas JJ, Jiménez E, López-Sánchez CA, Vega-Nieva DJ 513 (2020) Modeling of Aboveground Biomass with Landsat 8 OLI and Machine Learning in 514 Temperate Forests. Forests 11: 11 515 Lu D (2005) Aboveground biomass estimation using Landsat TM data in the Brazilian Amazon. 516 International Journal of Remote Sensing 26: 2509 – 2525 Electronic copy available at: https://ssrn.com/abstract=3954017 517 Lu D, Chen Q, Wang G, Liu L, Li G, Moran E (2016) A survey of remote sensing-based aboveground 518 biomass estimation methods in forest ecosystems. International Journal of Digital Earth 9: 63 – 519 105 520 MacLean DA, Wein RW (1977) Nutrient accumulation for postfire jack pine and hardwood succession 521 patterns in New Brunswick. Canadian Journal of Forest Research 7: 562 – 578 522 Marschner H (1991) Mechanisms of plant adaptation to acid soils. Plant and Soil 134: 1 – 20 523 Martin AR, Doraisami M, Thomas SC (2018) Global patterns in wood carbon concentration across the 524 world’s trees and forests. Nature Geoscience 11: 915 – 920 525 Meyer-Jacob C, Labaj AL, Paterson AM, Edwards BA, Keller W, Cummings BF, Smol JP (2020) Re- 526 browning of Sudbury (Ontario, Canada) lakes now approaches pre-acid deposition lake-water 527 dissolved organic carbon levels. Science of the Total Environment: 138347 528 Morrison IK (1973) Distribution of elements in aerial components of several natural jack pine stands in 529 northern Ontario. Canadian Journal of Forest Research 3: 170 – 179 530 Munford KE, Casamatta M, Basiliko N, Glasauer S, Mykytczuk NCS, Watmough SA (2021) Paper birch 531 (Betula papyrifera) nutrient resorption rates on nutrient-poor metal-contaminated soils and 532 mine tailings. Water, Air, & Soil Pollution 232: 33 533 Narendrula-Kotha R, and Nkongolo KK (2017). Microbial response to soil liming of damaged ecosystems 534 revealed by pyrosequencing and phospholipid fatty acid analyses. PloS one 12: e0168497. 535 Natural Resources Canada (NRCAN) (2014) Learn the facts: fuel consumption and CO 536 https://www.nrcan.gc.ca/sites/www.nrcan.gc.ca/files/oee/pdf/transportation/fuel-efficient- 537 technologies/autosmart_factsheet_6_e.pdf 538 Neumann M, Ukonmaanaho L, Johnson J, Benham S, Vesterdal L, Novotný R, Verstraeten A, Lundin L, 539 Thimonier A, Michopoulos P, Hasenauer H (2018) Quantifying carbon and nutrient input from Electronic copy available at: https://ssrn.com/abstract=3954017 540 litterfall in European forests using field observations and modeling. Global Biogeochemical 541 Cycles. 32: 784 – 798 542 Nkongolo KK, Spiers G, Beckett P, Narendrula R, Theriault G, Tran A, Kalubi KN (2013) Long-term effects 543 of liming on soil chemistry in stable and eroded upland areas in a mining region. Water, Air, & 544 Soil Pollution 224: 1618 545 Pardo LH, Duarte N, Miller EK, Robin-Abbot M (2005) Tree chemistry database (Version 1). USDA Forest 546 Service, Northeastern Research Station 547 Park A (2015) Carbon storage and stand conversion in a pine-dominated boreal forest landscape. Forest 548 Ecology and Management 340: 70 – 81 549 Pearson DAB, Pitblado JR (1995) Geological and geographical setting. Pages 5 – 15 In: Gunn J (ed) 550 Restoration and recovery of an industrial region. Springer, New York, USA 551 Pietrzykowski M, Daniels WL (2014) Estimation of carbon sequestration by pine (Pinus sylvestris L) 552 ecosystems developed on reforested post-mining sites in Poland on differing mine soil 553 substrates. Ecological Engineering 73: 209 – 218 554 Preston MD, Brummell ME, Smenderovac E, Rantala-Sykes B, Rumney RHM, Sherman G, Basiliko N, 555 Beckett P, Hebert M (2020) Tree restoration and ecosystem carbon storage in an acid and metal 556 impacted landscape: Chronosequence and resampling approaches. Forest Ecology and 557 Management 463: 118012 558 QGIS.org (2020) QGIS Geographic Information System http://www.qgis.org 559 Qi J, Chehbouni A, Huete AR, Kerr YH, & Sorooshian S (1994) A modified soil adjusted vegetation 560 index. Remote sensing of environment 48: 119 – 126 561 Rogelj J, Shindell D, Jiang K, Fifita S, Forster P, Ginzburg V, Handa C, Kheshgi H, Kobayashi S, Kriegler E, 562 Mundaca L, Séférian R, Vilariño MV (2018) Chapter 2: Mitigation pathways compatible with 563 1.5°C in the context of sustainable development. In: Masson-Delmotte V, Zhai P, Pörtner, HO, Electronic copy available at: https://ssrn.com/abstract=3954017 564 Roberts D, Skea J, Shukla PR, Pirani A, Moufouma-Okia W, Péan C, Pidcock R, Connors S, 565 Matthews JBR, Chen Y, Zhou X, Gomis MI, Lonnoy E, Maycock T, Tignor M, Waterfield T (eds) 566 Global Warming of 1.5°C. An IPCC Special Report on the impacts of global warming of 1.5°C 567 above pre-industrial levels and related global greenhouse gas emission pathways, in the context 568 of strengthening the global response to the threat of climate change, sustainable development, 569 and efforts to eradicate poverty. Intergovernmental Panel on Climate Change. 570 Rowe JS (1972) Forest regions of Canada. Canadian Forest Service, Department of Environment, Ottawa, 571 Canada 572 RStudio Team (2020) RStudio: Integrated Development for R http://www.rstudio.com/ 573 Rudolf PO (1990) Pinus resinosa Ait Red pine. Silvics of North America 1: 442 – 455 574 Rudolph TD, Laidly, PR (1990) Pinus banksiana Lamb Jack pine. Silvics of North America 1: 280 – 293 575 Rumney RHM, Preston MD, Jones T, Basiliko N, Gunn J (2021) Soil amendment improves carbon 576 sequestration by trees on severely damaged acid and metal landscape, but total storage remains 577 low. Forest Ecology and Management 483: 118896. 578 SARA Group (2009) Summary of Volume III: Ecological risk assessment. Sudbury Soils Study. SARA Group, 579 Guelph, Canada 580 Seidel D, Fleck S, Leuschner C, Hammett T (2011) Review of ground-based methods to measure the 581 distribution of biomass in forest canopies. Annals of Forest Science 68: 225 – 244 582 US Geological Survey (USGS) (2019) Landsat levels of processing details https://www.usgs.gov/core- 583 science-systems/nli/landsat/landsat-levels-processing 584 US Geological Survey (USGS) (2016) Landsat surface reflectance-derived spectral indices 585 https://www.usgs.gov/land-resources/nli/landsat/landsat-surface-reflectance-derived-spectral- 586 indices 587 US Geological Survey (USGS) (2020) EarthExplorer https://earthexplorer.usgs.gov/ Electronic copy available at: https://ssrn.com/abstract=3954017 588 Vegetation Enhancement Technical Advisory Committee (VETAC) (2020) 2020 Annual Report: 589 Regreening Program. VETAC, Sudbury, Canada 590 Winterhalder K (1995) Dynamics of plant communities and soils in revegetated ecosystems: a Sudbury 591 case study. Pages 173 – 182 In: Gunn J (ed) Restoration and recovery of an industrial region. 592 Springer, New York, USA 593 Winterhalder K (1996) Environmental degradation and rehabilitation of the landscape around Sudbury, a 594 major mining and smelting area. Environmental Reviews 4: 185 – 224 595 Wu C, Shen H, Wang K, Shen A, Deng J, Gan M (2016) Landsat imagery-based aboveground biomass 596 estimation and change investigation related to human activities. Sustainability 8: 159 597 Xiao X, Hollinger D, Aber J, Goltz M, Davidson EA, Zhang Q, Moore B (2004) Satellite-based modelling of 598 gross primary production in an evergreen needleleaf forest. Remote Sensing of the Environment 599 89: 519 – 534 600 Yanai RD (1992) Phosphorous budget of a 70-year-old northern hardwood forest. Biogeochemistry 17: 1 601 – 22 602 Zhang R, Zhuo X, Ouyang Z, Avitabile V, Qi J, Chen J, Giannico V (2019) Estimating aboveground biomass 603 in subtropical forests of China by integrating multisource remote sensing and ground data. 604 Remote Sensing of the Environment 232: 111341 Electronic copy available at: https://ssrn.com/abstract=3954017 612 Electronic copy available at: https://ssrn.com/abstract=3954017

Journal

AgricSciRN Subject Matter Journals – SSRN

Published: Nov 1, 2021

Keywords: forest restoration, degraded landscape, tree growth, forest nutrients, erosion

There are no references for this article.