add data citation

README.md

@@ -1,6 +1,6 @@
 ## Geographically aware estimates of remotely sensed water properties for Chesapeake Bay
 
-[![Paper DOI](https://img.shields.io/badge/Paper-XXXX/XXXX-blue.svg)](https://doi.org) [![Code DOI](https://img.shields.io/badge/Code-XXXX/XXXX-blue.svg)](https://doi.org) [![Data DOI](https://img.shields.io/badge/Data-XXXX/XXXX-blue.svg)](https://doi.org)
+[![Paper DOI](https://img.shields.io/badge/Paper-XXXX/XXXX-blue.svg)](https://doi.org) [![Code DOI](https://img.shields.io/badge/Code-XXXX/XXXX-blue.svg)](https://doi.org) [![Data DOI](https://img.shields.io/badge/Data-10.6084/m9.figshare.21578898-blue.svg)](https://doi.org/10.6084/m9.figshare.21578898)
 
 ### Products
 

manuscript/geowq.bib

@@ -406,6 +406,15 @@ @article{srichandanInterannualCyclonedrivenVariability2015
   langid = {english}
 }
 
+@misc{stachelekjemmachesapeake,
+  title = {Salinity, Temperature, and Turbidity Observations during 2000-2021 for {{Chesapeake Bay}}},
+  author = {Stachelek, Jemma},
+  year = {2022},
+  publisher = {{figshare}},
+  doi = {10.6084/m9.figshare.21578898},
+  copyright = {Creative Commons Attribution 4.0 International}
+}
+
 @article{turnerLongTermTrends2021,
   title = {Long-{{Term Trends}} in {{Chesapeake Bay Remote Sensing Reflectance}}: {{Implications}} for {{Water Clarity}}},
   shorttitle = {Long-{{Term Trends}} in {{Chesapeake Bay Remote Sensing Reflectance}}},

manuscript/manuscript.tex

@@ -63,7 +63,7 @@ \subsection{Data description}
 
 We supplemented observational data with a grid of “freshwater influence” (FWI) estimates where each grid cell represents the average of annual-flow weighted distance of that cell from the 5 major tributaries (Susquehanna, James, Pautexent, Choptank, and Potomac). Annual flow magnitudes were calculated from National Water Information System instantaneous data (sites: 01488110, 01491000, 01576000, 01654000, 01594440, 02037500, \cite{nwis2021usgs}). Rather than straight-line distances we used “in-water” distances calculated using the scikit-image Python package \cite{littleKrigingEstuariesCrow1997,vanderwaltScikitimageImageProcessing2014}. Our in-water distance calculations involved creating a cost–surface whereby land pixels had a very high cost relative to water pixels. Then the distance from any one cell to a tributary source was the accumulated travel “cost” as given by the scikit-image \texttt{route\_through\_array} function. We combined FWI estimates with the pixel-wise daily mean observations and regressed them in RF models against data from the \hl{1000m} resolution MODIS Aqua product, which provides \hl{9} bands of values for each pixel spanning from \hl{the following wavelength ranges (405-420, 438-448, 483-493, 526-536, 546-556, 662-672, 673-683, 743-753, and 862-877nm)} \cite{vermoteericMOD09GAMODISTerra2015}. \hl{MODIS data was pulled from Google Earth Engine (Asset: \texttt{MODIS/006/MYDOCGA}) using the \texttt{earthengine-api} Python package.}
 
-All code for data processing, model fitting, and model evaluation is available at Stachelek et al. (2023a) the pipeline for which is shown in Figure 1. All data used in the study are available in Stachelek et al. (2023b).
+All code for data processing, model fitting, and model evaluation is available at Stachelek et al. (2023a) the pipeline for which is shown in Figure 1. All data used in the study are available in \cite{stachelekjemmachesapeake}.
 
 \begin{figure}[ht!]
     \begin{center}